{"format": "ta-amps.seed-pack.verbs.v1", "pack_id": "evaluation-design", "status": "proposal_only", "tasks": [
{"queue": "conformance-seed-packs", "kind": "verb.problemize", "effect_class": "write-internal", "priority": 50, "idempotency_key": "seedpack:evaluation-design:design-evaluation:fc8645d987211134a4abd9080263804fef8f19faa0ffc10604f1fa177bad3eb3", "depends_on": [], "producer_surface": "surface:tessera", "visibility_scope": "surface_only", "allowed_surfaces": [], "target_cohorts": ["all"], "tags": ["category:verb.problemize", "pack:evaluation-design", "template:design-evaluation", "claimant:agent", "status:quarantined"], "brief": {"category_id": "verb.problemize", "contract": "ta-amps.seed-verb-template.v1", "pack": "evaluation-design", "pack_version": "0.1.0", "template_id": "evaluation-design.design-evaluation", "title": "Design the evaluation", "plain_summary": "Turn a question into an evaluation plan: arms, metrics, runs per arm, scoring and draft pass marks for a person to approve.", "outcome": "A plan, one arm noun per arm and draft pass marks, or no_change when an existing evaluation already answers the question.", "claimant": "agent", "inputs": {"pinned": [{"store": "seedpack.evaluation-design", "key": "evaluation-design.v0.1.0.method", "version": 1, "value_sha256": "436329f710d068a71b3e6f7252a8f4ff875bda0d6eac92424526cb5712be11db"}], "slots": {"question": null, "prior_evaluations": []}, "params": {"output_store": "seedpack.evaluation-design.outputs"}}, "slot_specs": {"question": {"category": "noun.seed_hypothesis", "required": true, "many": false, "description": "Pin of a noun stating the question and the decision it informs."}, "prior_evaluations": {"category": "noun.eval_run", "required": false, "many": true, "description": "Pins of earlier evaluation reports on the same question."}}, "param_specs": {"output_store": {"description": "Store for the produced nouns. The default is a placeholder; use the store your owner assigned.", "schema": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "required": true, "default": "seedpack.evaluation-design.outputs"}}, "input_schema": {"type": "object", "additionalProperties": false, "required": ["pinned", "slots", "params"], "properties": {"pinned": {"type": "array", "minItems": 1, "maxItems": 1, "items": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}, "slots": {"type": "object", "additionalProperties": false, "required": ["question"], "properties": {"question": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}, "prior_evaluations": {"type": "array", "minItems": 0, "maxItems": 10, "items": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}}}, "params": {"type": "object", "additionalProperties": false, "required": ["output_store"], "properties": {"output_store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}}}}}, "output_contract": {"visibility_scope": "surface_only", "allowed_surfaces": [], "write_mode": "create_only", "store_param": "output_store", "key_rule": "evaluation-design.<role>.<task_id>[.<n>]", "expected_version": 0, "produces": [{"name": "plan", "template": {"store": "seedpack.evaluation-design", "key": "evaluation-design.v0.1.0.template.evaluation-plan", "version": 1, "value_sha256": "6a0574ec2e905d769ae86efeed27fe4cb64936280ce5a0068aaf1400099d9718"}, "max": 1}, {"name": "arms", "template": {"store": "seedpack.evaluation-design", "key": "evaluation-design.v0.1.0.template.arm", "version": 1, "value_sha256": "d3a60352f0d3260cbd4b344b723a28966532a496d163460f90ed5522d56c4691"}, "max": 10}, {"name": "draft_pass_marks", "template": {"store": "seedpack.evaluation-design", "key": "evaluation-design.v0.1.0.template.pass-marks", "version": 1, "value_sha256": "c5c105e33276b5eb4bcfe772512d3df5fee7e189b69799c9bc356437046076e0"}, "max": 1}], "delivery": "Write produced nouns surface_only under your own authenticated surface, which is the requester's assigned surface, so the requester and that surface's reviewers can read them. Never use public_commons or allowlist other surfaces. Return every produced noun's pin in output.artifacts; add its value when the requester asks for inline delivery."}, "output_schema": {"type": "object", "additionalProperties": false, "required": ["outcome", "summary", "artifacts", "checks", "unknowns"], "properties": {"outcome": {"type": "string", "enum": ["plan_ready", "no_change", "unknown", "blocked"]}, "summary": {"type": "string", "minLength": 1, "maxLength": 1200}, "artifacts": {"type": "array", "maxItems": 12, "items": {"type": "object", "additionalProperties": false, "required": ["role", "ref"], "properties": {"role": {"type": "string", "minLength": 1, "maxLength": 64}, "ref": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}, "value": {"type": "object"}}}}, "checks": {"type": "array", "maxItems": 64, "items": {"type": "object", "additionalProperties": false, "required": ["id", "result"], "properties": {"id": {"type": "string", "minLength": 1, "maxLength": 80}, "result": {"type": "string", "enum": ["pass", "fail", "not_run"]}, "evidence": {"type": "string", "maxLength": 2000}}}}, "unknowns": {"type": "array", "maxItems": 32, "items": {"type": "object", "additionalProperties": false, "required": ["item", "reason"], "properties": {"item": {"type": "string", "minLength": 1, "maxLength": 300}, "reason": {"type": "string", "minLength": 1, "maxLength": 1000}, "would_resolve": {"type": "string", "maxLength": 1000}}}}, "no_change_reason": {"type": "string", "minLength": 1, "maxLength": 2000}, "blocked_question": {"type": "string", "minLength": 1, "maxLength": 1000}, "proposed_verbs": {"type": "array", "maxItems": 10, "items": {"type": "object", "additionalProperties": false, "required": ["kind", "title", "outcome", "why", "priority"], "properties": {"kind": {"type": "string", "pattern": "^verb\\.[a-z_]+$"}, "template_id": {"type": "string", "minLength": 1, "maxLength": 160}, "title": {"type": "string", "minLength": 1, "maxLength": 160}, "outcome": {"type": "string", "minLength": 1, "maxLength": 600}, "why": {"type": "string", "minLength": 1, "maxLength": 600}, "input_pins": {"type": "array", "maxItems": 16, "items": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}, "priority": {"type": "integer", "minimum": 1, "maximum": 3}}}}, "proposed_nouns": {"type": "array", "maxItems": 10, "items": {"type": "object", "additionalProperties": false, "required": ["category", "role", "title", "why"], "properties": {"category": {"type": "string", "pattern": "^noun\\.[a-z_]+$"}, "role": {"type": "string", "minLength": 1, "maxLength": 64}, "title": {"type": "string", "minLength": 1, "maxLength": 160}, "why": {"type": "string", "minLength": 1, "maxLength": 600}}}}, "run_notes": {"type": "object", "additionalProperties": false, "properties": {"went_right": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}, "went_wrong": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}, "next_steps": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}, "ideas": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}}}}}, "permitted_outcomes": ["plan_ready", "no_change", "unknown", "blocked"], "outcome_meanings": {"plan_ready": "The declared nouns were produced and every deterministic acceptance check passed.", "no_change": "The evidence does not justify producing or changing anything. Say why.", "unknown": "The question could not be answered with the permitted inputs. List what is missing and what would resolve it.", "blocked": "A required input, permission or capability is missing. Ask one precise question."}, "outcome_requirements": {"plan_ready": ["artifacts"], "no_change": ["no_change_reason"], "unknown": ["unknowns"], "blocked": ["blocked_question"]}, "acceptance_checks": [{"id": "plan-matches-template", "check": "The plan validates against template.evaluation-plan's schema.", "method": "deterministic"}, {"id": "has-baseline", "check": "Exactly one arm has role baseline, and runs_per_arm is at least 2.", "method": "deterministic"}, {"id": "arms-exact", "check": "Every arm noun names an exact model identifier and a settings source.", "method": "deterministic"}, {"id": "marks-drafted", "check": "The draft pass marks reference only arms and metrics in the plan.", "method": "deterministic"}, {"id": "answers-question", "check": "The arms and metrics can actually answer the stated question.", "method": "judgment"}], "creates": {"nouns": "The worker writes the nouns listed in output_contract.produces and returns their pins.", "verbs": "The worker proposes follow-up verbs in output.proposed_verbs, such as a breakdown into smaller verbs or a research follow-up, with the pins they need. Proposals are never enqueued by the worker; an authorized requester enqueues the ones worth doing."}, "guidance": {"required": ["Verify the pins. State the question and the decision it informs in one sentence each.", "Define a baseline and the fewest arms that answer the question, one change per arm where possible.", "Pick metrics from the method's menu and say why each matters to the decision.", "Draft pass marks with thresholds; they stay drafts until a person commits them.", "Estimate cost per arm from dated prices and label it an estimate.", "Propose a freeze-set verb, a commit-pass-marks verb for a person, and two run-arm verbs per arm to follow them.", "Return follow-up work the results justify as proposed_verbs and nouns worth creating later as proposed_nouns. Never enqueue them yourself; an authorized requester decides which to run."], "preferred": [], "suggested": [], "discovery_requested": false}, "required_capabilities": ["read permitted noun versions", "write candidate nouns in the requester's scope"], "bounds": {"max_attempts": 2, "max_elapsed_seconds": 1800}, "execution": {"reasoning": "authorized_worker", "model_use": "worker_allowance", "inference_allowance": "Existing authorized worker allowance only. This template grants no provider credit, spending authority or server inference.", "server_fallback": false, "maximum_automatically_enqueued_children": 0, "network": {"mode": "none", "detail": "No network access."}, "external_effects": "none"}, "review": {"required": "owning_policy", "self_approval_permitted": false, "publication_permitted": false, "admission_permitted": false}, "instantiation": {"status": "template", "fill_before_enqueue": ["question"], "rule": "Replace each slot with exact NounRef pins, set the params, then recompute the idempotency key with scripts/build_packs.py --instantiate. Enqueue only with your own grants in a queue your owner assigned."}, "execution_report": {"optional": true, "purpose": "Lets later runs of this template be fingerprinted and compared. It is worker reported, not proof.", "template": {"version": 1, "harness": {"name": "replace-with-harness-name", "version": "replace-with-exact-version"}, "model": {"provider": "replace-with-provider-or-none", "name": "replace-with-exact-model-or-none"}, "summaries": {"problem": "One sentence: what this run was asked to resolve.", "input": "One sentence: which pinned inputs were read.", "solution": "One sentence: what was produced or why the outcome was no_change, unknown or blocked."}, "contract": {"input_schema": {"type": "object", "additionalProperties": false, "required": ["pinned", "slots", "params"], "properties": {"pinned": {"type": "array", "minItems": 1, "maxItems": 1, "items": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}, "slots": {"type": "object", "additionalProperties": false, "required": ["question"], "properties": {"question": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}, "prior_evaluations": {"type": "array", "minItems": 0, "maxItems": 10, "items": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}}}, "params": {"type": "object", "additionalProperties": false, "required": ["output_store"], "properties": {"output_store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}}}}}, "output_schema": {"type": "object", "additionalProperties": false, "required": ["outcome", "summary", "artifacts", "checks", "unknowns"], "properties": {"outcome": {"type": "string", "enum": ["plan_ready", "no_change", "unknown", "blocked"]}, "summary": {"type": "string", "minLength": 1, "maxLength": 1200}, "artifacts": {"type": "array", "maxItems": 12, "items": {"type": "object", "additionalProperties": false, "required": ["role", "ref"], "properties": {"role": {"type": "string", "minLength": 1, "maxLength": 64}, "ref": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}, "value": {"type": "object"}}}}, "checks": {"type": "array", "maxItems": 64, "items": {"type": "object", "additionalProperties": false, "required": ["id", "result"], "properties": {"id": {"type": "string", "minLength": 1, "maxLength": 80}, "result": {"type": "string", "enum": ["pass", "fail", "not_run"]}, "evidence": {"type": "string", "maxLength": 2000}}}}, "unknowns": {"type": "array", "maxItems": 32, "items": {"type": "object", "additionalProperties": false, "required": ["item", "reason"], "properties": {"item": {"type": "string", "minLength": 1, "maxLength": 300}, "reason": {"type": "string", "minLength": 1, "maxLength": 1000}, "would_resolve": {"type": "string", "maxLength": 1000}}}}, "no_change_reason": {"type": "string", "minLength": 1, "maxLength": 2000}, "blocked_question": {"type": "string", "minLength": 1, "maxLength": 1000}, "proposed_verbs": {"type": "array", "maxItems": 10, "items": {"type": "object", "additionalProperties": false, "required": ["kind", "title", "outcome", "why", "priority"], "properties": {"kind": {"type": "string", "pattern": "^verb\\.[a-z_]+$"}, "template_id": {"type": "string", "minLength": 1, "maxLength": 160}, "title": {"type": "string", "minLength": 1, "maxLength": 160}, "outcome": {"type": "string", "minLength": 1, "maxLength": 600}, "why": {"type": "string", "minLength": 1, "maxLength": 600}, "input_pins": {"type": "array", "maxItems": 16, "items": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}, "priority": {"type": "integer", "minimum": 1, "maximum": 3}}}}, "proposed_nouns": {"type": "array", "maxItems": 10, "items": {"type": "object", "additionalProperties": false, "required": ["category", "role", "title", "why"], "properties": {"category": {"type": "string", "pattern": "^noun\\.[a-z_]+$"}, "role": {"type": "string", "minLength": 1, "maxLength": 64}, "title": {"type": "string", "minLength": 1, "maxLength": 160}, "why": {"type": "string", "minLength": 1, "maxLength": 600}}}}, "run_notes": {"type": "object", "additionalProperties": false, "properties": {"went_right": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}, "went_wrong": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}, "next_steps": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}, "ideas": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}}}}}}, "blocks": [{"kind": "error", "text": ""}], "artifact_refs": [], "usage": {"source": "harness_metadata", "input_tokens": null, "output_tokens": null, "reason": "Template only. Replace with measured counters or keep null with a reason."}}, "fill_rules": ["Copy contract.input_schema and contract.output_schema unchanged, so every run of this template shares one shape fingerprint.", "Add context_hash, environment_hash and toolchain_hash as lowercase SHA-256 digests when known; all five contract fields together allow exact matching.", "Name the exact harness version, provider and model. Deterministic or human work records the model name as none.", "Put what went wrong in an error block and in output.run_notes.went_wrong. The version-1 report has no fields for what went right, next steps or ideas, so those go in output.run_notes.", "Leave token counts null with a reason unless the harness or provider measured them."], "run_notes_field": "output.run_notes", "shape_sha256": "05cacbabf4906bbe03a524800b646e4f41f26c43d3773a1beb739a5920fa1092"}, "dimensions": {"determinism": 0.0, "evidence_grounding": 0.0, "token_efficiency": 0.0, "safety_compliance": 0.0, "verification_tier": "unverified"}, "dimension_basis": {"measured": false, "method": "unmeasured_floor", "reason": "Tessera requires these axes, so each is set to its lowest value and the tier is unverified. No quality is claimed. Replace them only with measured values and the evidence that produced them."}, "relations": [{"relation": "requires_noun", "target_id": "seedpack.evaluation-design/evaluation-design.v0.1.0.method", "target_category": "noun.skill_pack", "note": ""}, {"relation": "produces_noun", "target_id": "seedpack.evaluation-design/evaluation-design.v0.1.0.template.evaluation-plan", "target_category": "noun.benchmark_suite", "note": "output plan"}, {"relation": "produces_noun", "target_id": "seedpack.evaluation-design/evaluation-design.v0.1.0.template.arm", "target_category": "noun.skill_pack", "note": "output arms"}, {"relation": "produces_noun", "target_id": "seedpack.evaluation-design/evaluation-design.v0.1.0.template.pass-marks", "target_category": "noun.seed_hypothesis", "note": "output draft_pass_marks"}]}},
{"queue": "conformance-seed-packs", "kind": "verb.ingest", "effect_class": "write-internal", "priority": 50, "idempotency_key": "seedpack:evaluation-design:freeze-set:6fd5d6d81897e153a395f63977048ed9248b05e8bbdc0c3d128d5104e3f079e8", "depends_on": [], "producer_surface": "surface:tessera", "visibility_scope": "surface_only", "allowed_surfaces": [], "target_cohorts": ["all"], "tags": ["category:verb.ingest", "pack:evaluation-design", "template:freeze-set", "claimant:agent", "status:quarantined"], "brief": {"category_id": "verb.ingest", "contract": "ta-amps.seed-verb-template.v1", "pack": "evaluation-design", "pack_version": "0.1.0", "template_id": "evaluation-design.freeze-set", "title": "Freeze the case set", "plain_summary": "Hash every case and label into a manifest with opaque ids before any arm runs, and check that ids cannot leak labels.", "outcome": "A frozen-set manifest; labels stay readable only by the scorer.", "claimant": "agent", "inputs": {"pinned": [{"store": "seedpack.evaluation-design", "key": "evaluation-design.v0.1.0.method", "version": 1, "value_sha256": "436329f710d068a71b3e6f7252a8f4ff875bda0d6eac92424526cb5712be11db"}], "slots": {"plan": null, "case_source": null}, "params": {"output_store": "seedpack.evaluation-design.outputs"}}, "slot_specs": {"plan": {"category": "noun.benchmark_suite", "required": true, "many": false, "description": "Pin of the evaluation plan."}, "case_source": {"category": "noun.benchmark_suite", "required": true, "many": false, "description": "Pin of the generator output or case collection to freeze."}}, "param_specs": {"output_store": {"description": "Store for the produced nouns. The default is a placeholder; use the store your owner assigned.", "schema": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "required": true, "default": "seedpack.evaluation-design.outputs"}}, "input_schema": {"type": "object", "additionalProperties": false, "required": ["pinned", "slots", "params"], "properties": {"pinned": {"type": "array", "minItems": 1, "maxItems": 1, "items": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}, "slots": {"type": "object", "additionalProperties": false, "required": ["plan", "case_source"], "properties": {"plan": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}, "case_source": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}}, "params": {"type": "object", "additionalProperties": false, "required": ["output_store"], "properties": {"output_store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}}}}}, "output_contract": {"visibility_scope": "surface_only", "allowed_surfaces": [], "write_mode": "create_only", "store_param": "output_store", "key_rule": "evaluation-design.<role>.<task_id>[.<n>]", "expected_version": 0, "produces": [{"name": "manifest", "template": {"store": "seedpack.evaluation-design", "key": "evaluation-design.v0.1.0.template.frozen-set-manifest", "version": 1, "value_sha256": "1639cfc067a9995a94d5e134ef35f6e1f6053a89d123fe767c602e94a82fe77e"}, "max": 1}], "delivery": "Write produced nouns surface_only under your own authenticated surface, which is the requester's assigned surface, so the requester and that surface's reviewers can read them. Never use public_commons or allowlist other surfaces. Return every produced noun's pin in output.artifacts; add its value when the requester asks for inline delivery."}, "output_schema": {"type": "object", "additionalProperties": false, "required": ["outcome", "summary", "artifacts", "checks", "unknowns"], "properties": {"outcome": {"type": "string", "enum": ["set_frozen", "no_change", "unknown", "blocked"]}, "summary": {"type": "string", "minLength": 1, "maxLength": 1200}, "artifacts": {"type": "array", "maxItems": 1, "items": {"type": "object", "additionalProperties": false, "required": ["role", "ref"], "properties": {"role": {"type": "string", "minLength": 1, "maxLength": 64}, "ref": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}, "value": {"type": "object"}}}}, "checks": {"type": "array", "maxItems": 64, "items": {"type": "object", "additionalProperties": false, "required": ["id", "result"], "properties": {"id": {"type": "string", "minLength": 1, "maxLength": 80}, "result": {"type": "string", "enum": ["pass", "fail", "not_run"]}, "evidence": {"type": "string", "maxLength": 2000}}}}, "unknowns": {"type": "array", "maxItems": 32, "items": {"type": "object", "additionalProperties": false, "required": ["item", "reason"], "properties": {"item": {"type": "string", "minLength": 1, "maxLength": 300}, "reason": {"type": "string", "minLength": 1, "maxLength": 1000}, "would_resolve": {"type": "string", "maxLength": 1000}}}}, "no_change_reason": {"type": "string", "minLength": 1, "maxLength": 2000}, "blocked_question": {"type": "string", "minLength": 1, "maxLength": 1000}, "proposed_verbs": {"type": "array", "maxItems": 10, "items": {"type": "object", "additionalProperties": false, "required": ["kind", "title", "outcome", "why", "priority"], "properties": {"kind": {"type": "string", "pattern": "^verb\\.[a-z_]+$"}, "template_id": {"type": "string", "minLength": 1, "maxLength": 160}, "title": {"type": "string", "minLength": 1, "maxLength": 160}, "outcome": {"type": "string", "minLength": 1, "maxLength": 600}, "why": {"type": "string", "minLength": 1, "maxLength": 600}, "input_pins": {"type": "array", "maxItems": 16, "items": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}, "priority": {"type": "integer", "minimum": 1, "maximum": 3}}}}, "proposed_nouns": {"type": "array", "maxItems": 10, "items": {"type": "object", "additionalProperties": false, "required": ["category", "role", "title", "why"], "properties": {"category": {"type": "string", "pattern": "^noun\\.[a-z_]+$"}, "role": {"type": "string", "minLength": 1, "maxLength": 64}, "title": {"type": "string", "minLength": 1, "maxLength": 160}, "why": {"type": "string", "minLength": 1, "maxLength": 600}}}}, "run_notes": {"type": "object", "additionalProperties": false, "properties": {"went_right": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}, "went_wrong": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}, "next_steps": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}, "ideas": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}}}}}, "permitted_outcomes": ["set_frozen", "no_change", "unknown", "blocked"], "outcome_meanings": {"set_frozen": "The declared nouns were produced and every deterministic acceptance check passed.", "no_change": "The evidence does not justify producing or changing anything. Say why.", "unknown": "The question could not be answered with the permitted inputs. List what is missing and what would resolve it.", "blocked": "A required input, permission or capability is missing. Ask one precise question."}, "outcome_requirements": {"set_frozen": ["artifacts"], "no_change": ["no_change_reason"], "unknown": ["unknowns"], "blocked": ["blocked_question"]}, "acceptance_checks": [{"id": "manifest-matches-template", "check": "The manifest validates against template.frozen-set-manifest's schema.", "method": "deterministic"}, {"id": "ids-opaque", "check": "No case id contains a label value, and ids sorted alphabetically are not grouped by label.", "method": "deterministic"}, {"id": "digests-unique", "check": "No two cases share an input digest.", "method": "deterministic"}, {"id": "groups-respected", "check": "Every group sits entirely inside one split.", "method": "deterministic"}], "creates": {"nouns": "The worker writes the nouns listed in output_contract.produces and returns their pins.", "verbs": "The worker proposes follow-up verbs in output.proposed_verbs, such as a breakdown into smaller verbs or a research follow-up, with the pins they need. Proposals are never enqueued by the worker; an authorized requester enqueues the ones worth doing."}, "guidance": {"required": ["Verify the pins. Assign opaque case ids that carry no label, order or source hints.", "Compute SHA-256 digests of each case input and each label; store labels where only the scorer can read them.", "Put near-duplicates in one group and one split.", "Record the generator or source pin and the freeze time.", "Return follow-up work the results justify as proposed_verbs and nouns worth creating later as proposed_nouns. Never enqueue them yourself; an authorized requester decides which to run."], "preferred": [], "suggested": [], "discovery_requested": false}, "required_capabilities": ["read permitted noun versions", "write candidate nouns in the requester's scope"], "bounds": {"max_attempts": 2, "max_elapsed_seconds": 1200}, "execution": {"reasoning": "deterministic_tools", "model_use": "none", "inference_allowance": "Existing authorized worker allowance only. This template grants no provider credit, spending authority or server inference.", "server_fallback": false, "maximum_automatically_enqueued_children": 0, "network": {"mode": "none", "detail": "No network access."}, "external_effects": "none"}, "review": {"required": "owning_policy", "self_approval_permitted": false, "publication_permitted": false, "admission_permitted": false}, "instantiation": {"status": "template", "fill_before_enqueue": ["plan", "case_source"], "rule": "Replace each slot with exact NounRef pins, set the params, then recompute the idempotency key with scripts/build_packs.py --instantiate. Enqueue only with your own grants in a queue your owner assigned."}, "execution_report": {"optional": true, "purpose": "Lets later runs of this template be fingerprinted and compared. It is worker reported, not proof.", "template": {"version": 1, "harness": {"name": "replace-with-harness-name", "version": "replace-with-exact-version"}, "model": {"provider": "replace-with-provider-or-none", "name": "replace-with-exact-model-or-none"}, "summaries": {"problem": "One sentence: what this run was asked to resolve.", "input": "One sentence: which pinned inputs were read.", "solution": "One sentence: what was produced or why the outcome was no_change, unknown or blocked."}, "contract": {"input_schema": {"type": "object", "additionalProperties": false, "required": ["pinned", "slots", "params"], "properties": {"pinned": {"type": "array", "minItems": 1, "maxItems": 1, "items": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}, "slots": {"type": "object", "additionalProperties": false, "required": ["plan", "case_source"], "properties": {"plan": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}, "case_source": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}}, "params": {"type": "object", "additionalProperties": false, "required": ["output_store"], "properties": {"output_store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}}}}}, "output_schema": {"type": "object", "additionalProperties": false, "required": ["outcome", "summary", "artifacts", "checks", "unknowns"], "properties": {"outcome": {"type": "string", "enum": ["set_frozen", "no_change", "unknown", "blocked"]}, "summary": {"type": "string", "minLength": 1, "maxLength": 1200}, "artifacts": {"type": "array", "maxItems": 1, "items": {"type": "object", "additionalProperties": false, "required": ["role", "ref"], "properties": {"role": {"type": "string", "minLength": 1, "maxLength": 64}, "ref": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}, "value": {"type": "object"}}}}, "checks": {"type": "array", "maxItems": 64, "items": {"type": "object", "additionalProperties": false, "required": ["id", "result"], "properties": {"id": {"type": "string", "minLength": 1, "maxLength": 80}, "result": {"type": "string", "enum": ["pass", "fail", "not_run"]}, "evidence": {"type": "string", "maxLength": 2000}}}}, "unknowns": {"type": "array", "maxItems": 32, "items": {"type": "object", "additionalProperties": false, "required": ["item", "reason"], "properties": {"item": {"type": "string", "minLength": 1, "maxLength": 300}, "reason": {"type": "string", "minLength": 1, "maxLength": 1000}, "would_resolve": {"type": "string", "maxLength": 1000}}}}, "no_change_reason": {"type": "string", "minLength": 1, "maxLength": 2000}, "blocked_question": {"type": "string", "minLength": 1, "maxLength": 1000}, "proposed_verbs": {"type": "array", "maxItems": 10, "items": {"type": "object", "additionalProperties": false, "required": ["kind", "title", "outcome", "why", "priority"], "properties": {"kind": {"type": "string", "pattern": "^verb\\.[a-z_]+$"}, "template_id": {"type": "string", "minLength": 1, "maxLength": 160}, "title": {"type": "string", "minLength": 1, "maxLength": 160}, "outcome": {"type": "string", "minLength": 1, "maxLength": 600}, "why": {"type": "string", "minLength": 1, "maxLength": 600}, "input_pins": {"type": "array", "maxItems": 16, "items": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}, "priority": {"type": "integer", "minimum": 1, "maximum": 3}}}}, "proposed_nouns": {"type": "array", "maxItems": 10, "items": {"type": "object", "additionalProperties": false, "required": ["category", "role", "title", "why"], "properties": {"category": {"type": "string", "pattern": "^noun\\.[a-z_]+$"}, "role": {"type": "string", "minLength": 1, "maxLength": 64}, "title": {"type": "string", "minLength": 1, "maxLength": 160}, "why": {"type": "string", "minLength": 1, "maxLength": 600}}}}, "run_notes": {"type": "object", "additionalProperties": false, "properties": {"went_right": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}, "went_wrong": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}, "next_steps": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}, "ideas": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}}}}}}, "blocks": [{"kind": "error", "text": ""}], "artifact_refs": [], "usage": {"source": "harness_metadata", "input_tokens": null, "output_tokens": null, "reason": "Template only. Replace with measured counters or keep null with a reason."}}, "fill_rules": ["Copy contract.input_schema and contract.output_schema unchanged, so every run of this template shares one shape fingerprint.", "Add context_hash, environment_hash and toolchain_hash as lowercase SHA-256 digests when known; all five contract fields together allow exact matching.", "Name the exact harness version, provider and model. Deterministic or human work records the model name as none.", "Put what went wrong in an error block and in output.run_notes.went_wrong. The version-1 report has no fields for what went right, next steps or ideas, so those go in output.run_notes.", "Leave token counts null with a reason unless the harness or provider measured them."], "run_notes_field": "output.run_notes", "shape_sha256": "8761953a2cd2212151135553651b2b4a29b377ed4c4ee2b7485d5807a509f731"}, "dimensions": {"determinism": 0.0, "evidence_grounding": 0.0, "token_efficiency": 0.0, "safety_compliance": 0.0, "verification_tier": "unverified"}, "dimension_basis": {"measured": false, "method": "unmeasured_floor", "reason": "Tessera requires these axes, so each is set to its lowest value and the tier is unverified. No quality is claimed. Replace them only with measured values and the evidence that produced them."}, "relations": [{"relation": "requires_noun", "target_id": "seedpack.evaluation-design/evaluation-design.v0.1.0.method", "target_category": "noun.skill_pack", "note": ""}, {"relation": "produces_noun", "target_id": "seedpack.evaluation-design/evaluation-design.v0.1.0.template.frozen-set-manifest", "target_category": "noun.benchmark_suite", "note": "output manifest"}], "follows_templates": ["evaluation-design.design-evaluation"]}},
{"queue": "conformance-seed-packs", "kind": "verb.verify_gate", "effect_class": "write-internal", "priority": 50, "idempotency_key": "seedpack:evaluation-design:commit-pass-marks:16af60e0f070ed1db68e0555031c49e143cfd1a139754ed76ef728e7b17aa479", "depends_on": [], "producer_surface": "surface:tessera", "visibility_scope": "surface_only", "allowed_surfaces": [], "target_cohorts": ["all"], "tags": ["category:verb.verify_gate", "pack:evaluation-design", "template:commit-pass-marks", "claimant:human", "status:quarantined"], "brief": {"category_id": "verb.verify_gate", "contract": "ta-amps.seed-verb-template.v1", "pack": "evaluation-design", "pack_version": "0.1.0", "template_id": "evaluation-design.commit-pass-marks", "title": "Commit the pass marks", "plain_summary": "A person reads the draft pass marks, changes them if needed, and commits them before any scored run.", "outcome": "Committed, immutable pass marks, or no_change if the person decides the evaluation should not run.", "claimant": "human", "inputs": {"pinned": [{"store": "seedpack.evaluation-design", "key": "evaluation-design.v0.1.0.method", "version": 1, "value_sha256": "436329f710d068a71b3e6f7252a8f4ff875bda0d6eac92424526cb5712be11db"}], "slots": {"plan": null, "draft_pass_marks": null}, "params": {"output_store": "seedpack.evaluation-design.outputs"}}, "slot_specs": {"plan": {"category": "noun.benchmark_suite", "required": true, "many": false, "description": "Pin of the evaluation plan."}, "draft_pass_marks": {"category": "noun.seed_hypothesis", "required": true, "many": false, "description": "Pin of the draft pass marks."}}, "param_specs": {"output_store": {"description": "Store for the produced nouns. The default is a placeholder; use the store your owner assigned.", "schema": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "required": true, "default": "seedpack.evaluation-design.outputs"}}, "input_schema": {"type": "object", "additionalProperties": false, "required": ["pinned", "slots", "params"], "properties": {"pinned": {"type": "array", "minItems": 1, "maxItems": 1, "items": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}, "slots": {"type": "object", "additionalProperties": false, "required": ["plan", "draft_pass_marks"], "properties": {"plan": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}, "draft_pass_marks": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}}, "params": {"type": "object", "additionalProperties": false, "required": ["output_store"], "properties": {"output_store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}}}}}, "output_contract": {"visibility_scope": "surface_only", "allowed_surfaces": [], "write_mode": "create_only", "store_param": "output_store", "key_rule": "evaluation-design.<role>.<task_id>[.<n>]", "expected_version": 0, "produces": [{"name": "pass_marks", "template": {"store": "seedpack.evaluation-design", "key": "evaluation-design.v0.1.0.template.pass-marks", "version": 1, "value_sha256": "c5c105e33276b5eb4bcfe772512d3df5fee7e189b69799c9bc356437046076e0"}, "max": 1}], "delivery": "Write produced nouns surface_only under your own authenticated surface, which is the requester's assigned surface, so the requester and that surface's reviewers can read them. Never use public_commons or allowlist other surfaces. Return every produced noun's pin in output.artifacts; add its value when the requester asks for inline delivery."}, "output_schema": {"type": "object", "additionalProperties": false, "required": ["outcome", "summary", "artifacts", "checks", "unknowns"], "properties": {"outcome": {"type": "string", "enum": ["marks_committed", "no_change", "unknown", "blocked"]}, "summary": {"type": "string", "minLength": 1, "maxLength": 1200}, "artifacts": {"type": "array", "maxItems": 1, "items": {"type": "object", "additionalProperties": false, "required": ["role", "ref"], "properties": {"role": {"type": "string", "minLength": 1, "maxLength": 64}, "ref": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}, "value": {"type": "object"}}}}, "checks": {"type": "array", "maxItems": 64, "items": {"type": "object", "additionalProperties": false, "required": ["id", "result"], "properties": {"id": {"type": "string", "minLength": 1, "maxLength": 80}, "result": {"type": "string", "enum": ["pass", "fail", "not_run"]}, "evidence": {"type": "string", "maxLength": 2000}}}}, "unknowns": {"type": "array", "maxItems": 32, "items": {"type": "object", "additionalProperties": false, "required": ["item", "reason"], "properties": {"item": {"type": "string", "minLength": 1, "maxLength": 300}, "reason": {"type": "string", "minLength": 1, "maxLength": 1000}, "would_resolve": {"type": "string", "maxLength": 1000}}}}, "no_change_reason": {"type": "string", "minLength": 1, "maxLength": 2000}, "blocked_question": {"type": "string", "minLength": 1, "maxLength": 1000}, "proposed_verbs": {"type": "array", "maxItems": 10, "items": {"type": "object", "additionalProperties": false, "required": ["kind", "title", "outcome", "why", "priority"], "properties": {"kind": {"type": "string", "pattern": "^verb\\.[a-z_]+$"}, "template_id": {"type": "string", "minLength": 1, "maxLength": 160}, "title": {"type": "string", "minLength": 1, "maxLength": 160}, "outcome": {"type": "string", "minLength": 1, "maxLength": 600}, "why": {"type": "string", "minLength": 1, "maxLength": 600}, "input_pins": {"type": "array", "maxItems": 16, "items": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}, "priority": {"type": "integer", "minimum": 1, "maximum": 3}}}}, "proposed_nouns": {"type": "array", "maxItems": 10, "items": {"type": "object", "additionalProperties": false, "required": ["category", "role", "title", "why"], "properties": {"category": {"type": "string", "pattern": "^noun\\.[a-z_]+$"}, "role": {"type": "string", "minLength": 1, "maxLength": 64}, "title": {"type": "string", "minLength": 1, "maxLength": 160}, "why": {"type": "string", "minLength": 1, "maxLength": 600}}}}, "run_notes": {"type": "object", "additionalProperties": false, "properties": {"went_right": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}, "went_wrong": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}, "next_steps": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}, "ideas": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}}}}}, "permitted_outcomes": ["marks_committed", "no_change", "unknown", "blocked"], "outcome_meanings": {"marks_committed": "The declared nouns were produced and every deterministic acceptance check passed.", "no_change": "The evidence does not justify producing or changing anything. Say why.", "unknown": "The question could not be answered with the permitted inputs. List what is missing and what would resolve it.", "blocked": "A required input, permission or capability is missing. Ask one precise question."}, "outcome_requirements": {"marks_committed": ["artifacts"], "no_change": ["no_change_reason"], "unknown": ["unknowns"], "blocked": ["blocked_question"]}, "acceptance_checks": [{"id": "marks-match-template", "check": "The committed marks validate against template.pass-marks's schema with immutable true.", "method": "deterministic"}, {"id": "before-runs", "check": "No run record for this evaluation existed when the marks were committed.", "method": "deterministic"}, {"id": "committed-by-person", "check": "committed_by is human and the submitting key belongs to a person.", "method": "deterministic"}], "creates": {"nouns": "The worker writes the nouns listed in output_contract.produces and returns their pins.", "verbs": "The worker proposes follow-up verbs in output.proposed_verbs, such as a breakdown into smaller verbs or a research follow-up, with the pins they need. Proposals are never enqueued by the worker; an authorized requester enqueues the ones worth doing."}, "guidance": {"required": ["Read the plan and the draft marks. Each mark should be something you would act on if it were met or missed.", "Change thresholds if needed, then commit. After this, the marks never change for this evaluation version.", "Commit before any run-arm task for this evaluation is claimed.", "Return follow-up work the results justify as proposed_verbs and nouns worth creating later as proposed_nouns. Never enqueue them yourself; an authorized requester decides which to run."], "preferred": [], "suggested": [], "discovery_requested": false}, "required_capabilities": ["read permitted noun versions", "write candidate nouns in the requester's scope"], "bounds": {"max_attempts": 2, "max_elapsed_seconds": 604800}, "execution": {"reasoning": "human", "model_use": "none", "inference_allowance": "Existing authorized worker allowance only. This template grants no provider credit, spending authority or server inference.", "server_fallback": false, "maximum_automatically_enqueued_children": 0, "network": {"mode": "none", "detail": "No network access."}, "external_effects": "none"}, "review": {"required": "owning_policy", "self_approval_permitted": false, "publication_permitted": false, "admission_permitted": false}, "instantiation": {"status": "template", "fill_before_enqueue": ["plan", "draft_pass_marks"], "rule": "Replace each slot with exact NounRef pins, set the params, then recompute the idempotency key with scripts/build_packs.py --instantiate. Enqueue only with your own grants in a queue your owner assigned."}, "execution_report": {"optional": true, "purpose": "Lets later runs of this template be fingerprinted and compared. It is worker reported, not proof.", "template": {"version": 1, "harness": {"name": "replace-with-harness-name", "version": "replace-with-exact-version"}, "model": {"provider": "replace-with-provider-or-none", "name": "replace-with-exact-model-or-none"}, "summaries": {"problem": "One sentence: what this run was asked to resolve.", "input": "One sentence: which pinned inputs were read.", "solution": "One sentence: what was produced or why the outcome was no_change, unknown or blocked."}, "contract": {"input_schema": {"type": "object", "additionalProperties": false, "required": ["pinned", "slots", "params"], "properties": {"pinned": {"type": "array", "minItems": 1, "maxItems": 1, "items": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}, "slots": {"type": "object", "additionalProperties": false, "required": ["plan", "draft_pass_marks"], "properties": {"plan": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}, "draft_pass_marks": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}}, "params": {"type": "object", "additionalProperties": false, "required": ["output_store"], "properties": {"output_store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}}}}}, "output_schema": {"type": "object", "additionalProperties": false, "required": ["outcome", "summary", "artifacts", "checks", "unknowns"], "properties": {"outcome": {"type": "string", "enum": ["marks_committed", "no_change", "unknown", "blocked"]}, "summary": {"type": "string", "minLength": 1, "maxLength": 1200}, "artifacts": {"type": "array", "maxItems": 1, "items": {"type": "object", "additionalProperties": false, "required": ["role", "ref"], "properties": {"role": {"type": "string", "minLength": 1, "maxLength": 64}, "ref": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}, "value": {"type": "object"}}}}, "checks": {"type": "array", "maxItems": 64, "items": {"type": "object", "additionalProperties": false, "required": ["id", "result"], "properties": {"id": {"type": "string", "minLength": 1, "maxLength": 80}, "result": {"type": "string", "enum": ["pass", "fail", "not_run"]}, "evidence": {"type": "string", "maxLength": 2000}}}}, "unknowns": {"type": "array", "maxItems": 32, "items": {"type": "object", "additionalProperties": false, "required": ["item", "reason"], "properties": {"item": {"type": "string", "minLength": 1, "maxLength": 300}, "reason": {"type": "string", "minLength": 1, "maxLength": 1000}, "would_resolve": {"type": "string", "maxLength": 1000}}}}, "no_change_reason": {"type": "string", "minLength": 1, "maxLength": 2000}, "blocked_question": {"type": "string", "minLength": 1, "maxLength": 1000}, "proposed_verbs": {"type": "array", "maxItems": 10, "items": {"type": "object", "additionalProperties": false, "required": ["kind", "title", "outcome", "why", "priority"], "properties": {"kind": {"type": "string", "pattern": "^verb\\.[a-z_]+$"}, "template_id": {"type": "string", "minLength": 1, "maxLength": 160}, "title": {"type": "string", "minLength": 1, "maxLength": 160}, "outcome": {"type": "string", "minLength": 1, "maxLength": 600}, "why": {"type": "string", "minLength": 1, "maxLength": 600}, "input_pins": {"type": "array", "maxItems": 16, "items": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}, "priority": {"type": "integer", "minimum": 1, "maximum": 3}}}}, "proposed_nouns": {"type": "array", "maxItems": 10, "items": {"type": "object", "additionalProperties": false, "required": ["category", "role", "title", "why"], "properties": {"category": {"type": "string", "pattern": "^noun\\.[a-z_]+$"}, "role": {"type": "string", "minLength": 1, "maxLength": 64}, "title": {"type": "string", "minLength": 1, "maxLength": 160}, "why": {"type": "string", "minLength": 1, "maxLength": 600}}}}, "run_notes": {"type": "object", "additionalProperties": false, "properties": {"went_right": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}, "went_wrong": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}, "next_steps": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}, "ideas": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}}}}}}, "blocks": [{"kind": "error", "text": ""}], "artifact_refs": [], "usage": {"source": "harness_metadata", "input_tokens": null, "output_tokens": null, "reason": "Template only. Replace with measured counters or keep null with a reason."}}, "fill_rules": ["Copy contract.input_schema and contract.output_schema unchanged, so every run of this template shares one shape fingerprint.", "Add context_hash, environment_hash and toolchain_hash as lowercase SHA-256 digests when known; all five contract fields together allow exact matching.", "Name the exact harness version, provider and model. Deterministic or human work records the model name as none.", "Put what went wrong in an error block and in output.run_notes.went_wrong. The version-1 report has no fields for what went right, next steps or ideas, so those go in output.run_notes.", "Leave token counts null with a reason unless the harness or provider measured them."], "run_notes_field": "output.run_notes", "shape_sha256": "99c528d769cb3a20fd2f726fa21ddf5b7f0992718c25613d7065a22acec7ac52"}, "dimensions": {"determinism": 0.0, "evidence_grounding": 0.0, "token_efficiency": 0.0, "safety_compliance": 0.0, "verification_tier": "unverified"}, "dimension_basis": {"measured": false, "method": "unmeasured_floor", "reason": "Tessera requires these axes, so each is set to its lowest value and the tier is unverified. No quality is claimed. Replace them only with measured values and the evidence that produced them."}, "relations": [{"relation": "requires_noun", "target_id": "seedpack.evaluation-design/evaluation-design.v0.1.0.method", "target_category": "noun.skill_pack", "note": ""}, {"relation": "produces_noun", "target_id": "seedpack.evaluation-design/evaluation-design.v0.1.0.template.pass-marks", "target_category": "noun.seed_hypothesis", "note": "output pass_marks"}], "follows_templates": ["evaluation-design.design-evaluation"]}},
{"queue": "conformance-seed-packs", "kind": "verb.benchmark_harness", "effect_class": "write-internal", "priority": 50, "idempotency_key": "seedpack:evaluation-design:run-arm:0ac25f1db9b9894b70e7687f2f3114b76899fe620e905b94d80a300d3e259307", "depends_on": [], "producer_surface": "surface:tessera", "visibility_scope": "surface_only", "allowed_surfaces": [], "target_cohorts": ["all"], "tags": ["category:verb.benchmark_harness", "pack:evaluation-design", "template:run-arm", "claimant:agent", "status:quarantined"], "brief": {"category_id": "verb.benchmark_harness", "contract": "ta-amps.seed-verb-template.v1", "pack": "evaluation-design", "pack_version": "0.1.0", "template_id": "evaluation-design.run-arm", "title": "Run one arm once", "plain_summary": "Run one arm over the whole frozen set exactly as defined and record every prediction, error, cost and latency.", "outcome": "A run record. Each arm runs twice, as two separate attempts with run_index 1 and 2.", "claimant": "agent", "inputs": {"pinned": [{"store": "seedpack.evaluation-design", "key": "evaluation-design.v0.1.0.method", "version": 1, "value_sha256": "436329f710d068a71b3e6f7252a8f4ff875bda0d6eac92424526cb5712be11db"}], "slots": {"plan": null, "arm": null, "manifest": null, "pass_marks": null}, "params": {"run_index": null, "output_store": "seedpack.evaluation-design.outputs"}}, "slot_specs": {"plan": {"category": "noun.benchmark_suite", "required": true, "many": false, "description": "Pin of the evaluation plan."}, "arm": {"category": "noun.skill_pack", "required": true, "many": false, "description": "Pin of the arm to run."}, "manifest": {"category": "noun.benchmark_suite", "required": true, "many": false, "description": "Pin of the frozen-set manifest."}, "pass_marks": {"category": "noun.seed_hypothesis", "required": true, "many": false, "description": "Pin of the committed pass marks."}}, "param_specs": {"run_index": {"description": "Which run of this arm, starting at 1.", "schema": {"type": "integer", "minimum": 1, "maximum": 5}, "required": true}, "output_store": {"description": "Store for the produced nouns. The default is a placeholder; use the store your owner assigned.", "schema": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "required": true, "default": "seedpack.evaluation-design.outputs"}}, "input_schema": {"type": "object", "additionalProperties": false, "required": ["pinned", "slots", "params"], "properties": {"pinned": {"type": "array", "minItems": 1, "maxItems": 1, "items": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}, "slots": {"type": "object", "additionalProperties": false, "required": ["plan", "arm", "manifest", "pass_marks"], "properties": {"plan": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}, "arm": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}, "manifest": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}, "pass_marks": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}}, "params": {"type": "object", "additionalProperties": false, "required": ["run_index", "output_store"], "properties": {"run_index": {"type": "integer", "minimum": 1, "maximum": 5}, "output_store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}}}}}, "output_contract": {"visibility_scope": "surface_only", "allowed_surfaces": [], "write_mode": "create_only", "store_param": "output_store", "key_rule": "evaluation-design.<role>.<task_id>[.<n>]", "expected_version": 0, "produces": [{"name": "run_record", "template": {"store": "seedpack.evaluation-design", "key": "evaluation-design.v0.1.0.template.run-record", "version": 1, "value_sha256": "4351a2712399c92113b08bc2f9fa75211e259332b84cd4e63479fa701ead40eb"}, "max": 1}], "delivery": "Write produced nouns surface_only under your own authenticated surface, which is the requester's assigned surface, so the requester and that surface's reviewers can read them. Never use public_commons or allowlist other surfaces. Return every produced noun's pin in output.artifacts; add its value when the requester asks for inline delivery."}, "output_schema": {"type": "object", "additionalProperties": false, "required": ["outcome", "summary", "artifacts", "checks", "unknowns"], "properties": {"outcome": {"type": "string", "enum": ["run_recorded", "no_change", "unknown", "blocked"]}, "summary": {"type": "string", "minLength": 1, "maxLength": 1200}, "artifacts": {"type": "array", "maxItems": 1, "items": {"type": "object", "additionalProperties": false, "required": ["role", "ref"], "properties": {"role": {"type": "string", "minLength": 1, "maxLength": 64}, "ref": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}, "value": {"type": "object"}}}}, "checks": {"type": "array", "maxItems": 64, "items": {"type": "object", "additionalProperties": false, "required": ["id", "result"], "properties": {"id": {"type": "string", "minLength": 1, "maxLength": 80}, "result": {"type": "string", "enum": ["pass", "fail", "not_run"]}, "evidence": {"type": "string", "maxLength": 2000}}}}, "unknowns": {"type": "array", "maxItems": 32, "items": {"type": "object", "additionalProperties": false, "required": ["item", "reason"], "properties": {"item": {"type": "string", "minLength": 1, "maxLength": 300}, "reason": {"type": "string", "minLength": 1, "maxLength": 1000}, "would_resolve": {"type": "string", "maxLength": 1000}}}}, "no_change_reason": {"type": "string", "minLength": 1, "maxLength": 2000}, "blocked_question": {"type": "string", "minLength": 1, "maxLength": 1000}, "proposed_verbs": {"type": "array", "maxItems": 10, "items": {"type": "object", "additionalProperties": false, "required": ["kind", "title", "outcome", "why", "priority"], "properties": {"kind": {"type": "string", "pattern": "^verb\\.[a-z_]+$"}, "template_id": {"type": "string", "minLength": 1, "maxLength": 160}, "title": {"type": "string", "minLength": 1, "maxLength": 160}, "outcome": {"type": "string", "minLength": 1, "maxLength": 600}, "why": {"type": "string", "minLength": 1, "maxLength": 600}, "input_pins": {"type": "array", "maxItems": 16, "items": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}, "priority": {"type": "integer", "minimum": 1, "maximum": 3}}}}, "proposed_nouns": {"type": "array", "maxItems": 10, "items": {"type": "object", "additionalProperties": false, "required": ["category", "role", "title", "why"], "properties": {"category": {"type": "string", "pattern": "^noun\\.[a-z_]+$"}, "role": {"type": "string", "minLength": 1, "maxLength": 64}, "title": {"type": "string", "minLength": 1, "maxLength": 160}, "why": {"type": "string", "minLength": 1, "maxLength": 600}}}}, "run_notes": {"type": "object", "additionalProperties": false, "properties": {"went_right": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}, "went_wrong": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}, "next_steps": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}, "ideas": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}}}}}, "permitted_outcomes": ["run_recorded", "no_change", "unknown", "blocked"], "outcome_meanings": {"run_recorded": "The declared nouns were produced and every deterministic acceptance check passed.", "no_change": "The evidence does not justify producing or changing anything. Say why.", "unknown": "The question could not be answered with the permitted inputs. List what is missing and what would resolve it.", "blocked": "A required input, permission or capability is missing. Ask one precise question."}, "outcome_requirements": {"run_recorded": ["artifacts"], "no_change": ["no_change_reason"], "unknown": ["unknowns"], "blocked": ["blocked_question"]}, "acceptance_checks": [{"id": "record-matches-template", "check": "The run record validates against template.run-record's schema.", "method": "deterministic"}, {"id": "all-cases", "check": "The record has exactly one entry per case in the manifest.", "method": "deterministic"}, {"id": "marks-predate", "check": "The pass marks were committed before this task was claimed.", "method": "deterministic"}, {"id": "arm-faithful", "check": "The run used the arm's exact model and settings.", "method": "judgment"}], "creates": {"nouns": "The worker writes the nouns listed in output_contract.produces and returns their pins.", "verbs": "The worker proposes follow-up verbs in output.proposed_verbs, such as a breakdown into smaller verbs or a research follow-up, with the pins they need. Proposals are never enqueued by the worker; an authorized requester enqueues the ones worth doing."}, "guidance": {"required": ["Verify every pin, and that the pass marks are committed and predate this claim; otherwise return blocked.", "Run the arm exactly as its noun defines it: same model, prompt, tools, context and settings. Keep the caller's settings unless the arm says otherwise.", "Never show case labels to the arm. Answer every case in the set; record errors and timeouts as they happened.", "Record cost and latency per case from actual usage; unknown values stay null with a reason.", "Stop at the arm's budget and record how far the run got.", "Return follow-up work the results justify as proposed_verbs and nouns worth creating later as proposed_nouns. Never enqueue them yourself; an authorized requester decides which to run."], "preferred": [], "suggested": [], "discovery_requested": false}, "required_capabilities": ["read permitted noun versions", "the arm's provider access within the worker's own allowance", "write candidate nouns in the requester's scope"], "bounds": {"max_attempts": 2, "max_elapsed_seconds": 14400}, "execution": {"reasoning": "authorized_worker", "model_use": "worker_allowance", "inference_allowance": "Existing authorized worker allowance only. This template grants no provider credit, spending authority or server inference.", "server_fallback": false, "maximum_automatically_enqueued_children": 0, "network": {"mode": "provider_destination", "detail": "Only the arm's declared provider destination."}, "external_effects": "none"}, "review": {"required": "owning_policy", "self_approval_permitted": false, "publication_permitted": false, "admission_permitted": false}, "instantiation": {"status": "template", "fill_before_enqueue": ["plan", "arm", "manifest", "pass_marks"], "rule": "Replace each slot with exact NounRef pins, set the params, then recompute the idempotency key with scripts/build_packs.py --instantiate. Enqueue only with your own grants in a queue your owner assigned."}, "execution_report": {"optional": true, "purpose": "Lets later runs of this template be fingerprinted and compared. It is worker reported, not proof.", "template": {"version": 1, "harness": {"name": "replace-with-harness-name", "version": "replace-with-exact-version"}, "model": {"provider": "replace-with-provider-or-none", "name": "replace-with-exact-model-or-none"}, "summaries": {"problem": "One sentence: what this run was asked to resolve.", "input": "One sentence: which pinned inputs were read.", "solution": "One sentence: what was produced or why the outcome was no_change, unknown or blocked."}, "contract": {"input_schema": {"type": "object", "additionalProperties": false, "required": ["pinned", "slots", "params"], "properties": {"pinned": {"type": "array", "minItems": 1, "maxItems": 1, "items": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}, "slots": {"type": "object", "additionalProperties": false, "required": ["plan", "arm", "manifest", "pass_marks"], "properties": {"plan": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}, "arm": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}, "manifest": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}, "pass_marks": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}}, "params": {"type": "object", "additionalProperties": false, "required": ["run_index", "output_store"], "properties": {"run_index": {"type": "integer", "minimum": 1, "maximum": 5}, "output_store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}}}}}, "output_schema": {"type": "object", "additionalProperties": false, "required": ["outcome", "summary", "artifacts", "checks", "unknowns"], "properties": {"outcome": {"type": "string", "enum": ["run_recorded", "no_change", "unknown", "blocked"]}, "summary": {"type": "string", "minLength": 1, "maxLength": 1200}, "artifacts": {"type": "array", "maxItems": 1, "items": {"type": "object", "additionalProperties": false, "required": ["role", "ref"], "properties": {"role": {"type": "string", "minLength": 1, "maxLength": 64}, "ref": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}, "value": {"type": "object"}}}}, "checks": {"type": "array", "maxItems": 64, "items": {"type": "object", "additionalProperties": false, "required": ["id", "result"], "properties": {"id": {"type": "string", "minLength": 1, "maxLength": 80}, "result": {"type": "string", "enum": ["pass", "fail", "not_run"]}, "evidence": {"type": "string", "maxLength": 2000}}}}, "unknowns": {"type": "array", "maxItems": 32, "items": {"type": "object", "additionalProperties": false, "required": ["item", "reason"], "properties": {"item": {"type": "string", "minLength": 1, "maxLength": 300}, "reason": {"type": "string", "minLength": 1, "maxLength": 1000}, "would_resolve": {"type": "string", "maxLength": 1000}}}}, "no_change_reason": {"type": "string", "minLength": 1, "maxLength": 2000}, "blocked_question": {"type": "string", "minLength": 1, "maxLength": 1000}, "proposed_verbs": {"type": "array", "maxItems": 10, "items": {"type": "object", "additionalProperties": false, "required": ["kind", "title", "outcome", "why", "priority"], "properties": {"kind": {"type": "string", "pattern": "^verb\\.[a-z_]+$"}, "template_id": {"type": "string", "minLength": 1, "maxLength": 160}, "title": {"type": "string", "minLength": 1, "maxLength": 160}, "outcome": {"type": "string", "minLength": 1, "maxLength": 600}, "why": {"type": "string", "minLength": 1, "maxLength": 600}, "input_pins": {"type": "array", "maxItems": 16, "items": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}, "priority": {"type": "integer", "minimum": 1, "maximum": 3}}}}, "proposed_nouns": {"type": "array", "maxItems": 10, "items": {"type": "object", "additionalProperties": false, "required": ["category", "role", "title", "why"], "properties": {"category": {"type": "string", "pattern": "^noun\\.[a-z_]+$"}, "role": {"type": "string", "minLength": 1, "maxLength": 64}, "title": {"type": "string", "minLength": 1, "maxLength": 160}, "why": {"type": "string", "minLength": 1, "maxLength": 600}}}}, "run_notes": {"type": "object", "additionalProperties": false, "properties": {"went_right": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}, "went_wrong": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}, "next_steps": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}, "ideas": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}}}}}}, "blocks": [{"kind": "error", "text": ""}], "artifact_refs": [], "usage": {"source": "harness_metadata", "input_tokens": null, "output_tokens": null, "reason": "Template only. Replace with measured counters or keep null with a reason."}}, "fill_rules": ["Copy contract.input_schema and contract.output_schema unchanged, so every run of this template shares one shape fingerprint.", "Add context_hash, environment_hash and toolchain_hash as lowercase SHA-256 digests when known; all five contract fields together allow exact matching.", "Name the exact harness version, provider and model. Deterministic or human work records the model name as none.", "Put what went wrong in an error block and in output.run_notes.went_wrong. The version-1 report has no fields for what went right, next steps or ideas, so those go in output.run_notes.", "Leave token counts null with a reason unless the harness or provider measured them."], "run_notes_field": "output.run_notes", "shape_sha256": "1089cb573f6c6525f2329e5b4626ea4f4dbc21ff2cca225fbcb23cd2deced251"}, "dimensions": {"determinism": 0.0, "evidence_grounding": 0.0, "token_efficiency": 0.0, "safety_compliance": 0.0, "verification_tier": "unverified"}, "dimension_basis": {"measured": false, "method": "unmeasured_floor", "reason": "Tessera requires these axes, so each is set to its lowest value and the tier is unverified. No quality is claimed. Replace them only with measured values and the evidence that produced them."}, "relations": [{"relation": "requires_noun", "target_id": "seedpack.evaluation-design/evaluation-design.v0.1.0.method", "target_category": "noun.skill_pack", "note": ""}, {"relation": "produces_noun", "target_id": "seedpack.evaluation-design/evaluation-design.v0.1.0.template.run-record", "target_category": "noun.eval_run", "note": "output run_record"}], "follows_templates": ["evaluation-design.commit-pass-marks", "evaluation-design.freeze-set"]}},
{"queue": "conformance-seed-packs", "kind": "verb.verify_gate", "effect_class": "write-internal", "priority": 50, "idempotency_key": "seedpack:evaluation-design:score-run:7f5d96114893a9eb2a36c258201329707b8a677c947f3b063341a880c8d17a59", "depends_on": [], "producer_surface": "surface:tessera", "visibility_scope": "surface_only", "allowed_surfaces": [], "target_cohorts": ["all"], "tags": ["category:verb.verify_gate", "pack:evaluation-design", "template:score-run", "claimant:agent", "status:quarantined"], "brief": {"category_id": "verb.verify_gate", "contract": "ta-amps.seed-verb-template.v1", "pack": "evaluation-design", "pack_version": "0.1.0", "template_id": "evaluation-design.score-run", "title": "Score one run", "plain_summary": "Score one run record by exact match against the frozen labels with the deterministic scoring procedure.", "outcome": "A scored run with metrics; the scorer is the only worker that reads labels.", "claimant": "agent", "inputs": {"pinned": [{"store": "seedpack.evaluation-design", "key": "evaluation-design.v0.1.0.scoring.exact-match", "version": 1, "value_sha256": "677024a78f60eef2c8391d61e0ca0a77e71a3443477b6bd6d64d0d400a68ae27"}, {"store": "seedpack.evaluation-design", "key": "evaluation-design.v0.1.0.method", "version": 1, "value_sha256": "436329f710d068a71b3e6f7252a8f4ff875bda0d6eac92424526cb5712be11db"}, {"store": "seedpack.evaluation-design", "key": "evaluation-design.v0.1.0.example.worked", "version": 1, "value_sha256": "68a2a02056104e275efaf2ea42233bf1276c33c9c9d4b18d5429fa7f1c859849"}], "slots": {"run_record": null, "manifest": null, "pass_marks": null}, "params": {"output_store": "seedpack.evaluation-design.outputs"}}, "slot_specs": {"run_record": {"category": "noun.eval_run", "required": true, "many": false, "description": "Pin of the run record."}, "manifest": {"category": "noun.benchmark_suite", "required": true, "many": false, "description": "Pin of the frozen-set manifest."}, "pass_marks": {"category": "noun.seed_hypothesis", "required": true, "many": false, "description": "Pin of the committed pass marks."}}, "param_specs": {"output_store": {"description": "Store for the produced nouns. The default is a placeholder; use the store your owner assigned.", "schema": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "required": true, "default": "seedpack.evaluation-design.outputs"}}, "input_schema": {"type": "object", "additionalProperties": false, "required": ["pinned", "slots", "params"], "properties": {"pinned": {"type": "array", "minItems": 3, "maxItems": 3, "items": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}, "slots": {"type": "object", "additionalProperties": false, "required": ["run_record", "manifest", "pass_marks"], "properties": {"run_record": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}, "manifest": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}, "pass_marks": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}}, "params": {"type": "object", "additionalProperties": false, "required": ["output_store"], "properties": {"output_store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}}}}}, "output_contract": {"visibility_scope": "surface_only", "allowed_surfaces": [], "write_mode": "create_only", "store_param": "output_store", "key_rule": "evaluation-design.<role>.<task_id>[.<n>]", "expected_version": 0, "produces": [{"name": "scored_run", "template": {"store": "seedpack.evaluation-design", "key": "evaluation-design.v0.1.0.template.run-record", "version": 1, "value_sha256": "4351a2712399c92113b08bc2f9fa75211e259332b84cd4e63479fa701ead40eb"}, "max": 1}], "delivery": "Write produced nouns surface_only under your own authenticated surface, which is the requester's assigned surface, so the requester and that surface's reviewers can read them. Never use public_commons or allowlist other surfaces. Return every produced noun's pin in output.artifacts; add its value when the requester asks for inline delivery."}, "output_schema": {"type": "object", "additionalProperties": false, "required": ["outcome", "summary", "artifacts", "checks", "unknowns"], "properties": {"outcome": {"type": "string", "enum": ["run_scored", "no_change", "unknown", "blocked"]}, "summary": {"type": "string", "minLength": 1, "maxLength": 1200}, "artifacts": {"type": "array", "maxItems": 1, "items": {"type": "object", "additionalProperties": false, "required": ["role", "ref"], "properties": {"role": {"type": "string", "minLength": 1, "maxLength": 64}, "ref": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}, "value": {"type": "object"}}}}, "checks": {"type": "array", "maxItems": 64, "items": {"type": "object", "additionalProperties": false, "required": ["id", "result"], "properties": {"id": {"type": "string", "minLength": 1, "maxLength": 80}, "result": {"type": "string", "enum": ["pass", "fail", "not_run"]}, "evidence": {"type": "string", "maxLength": 2000}}}}, "unknowns": {"type": "array", "maxItems": 32, "items": {"type": "object", "additionalProperties": false, "required": ["item", "reason"], "properties": {"item": {"type": "string", "minLength": 1, "maxLength": 300}, "reason": {"type": "string", "minLength": 1, "maxLength": 1000}, "would_resolve": {"type": "string", "maxLength": 1000}}}}, "no_change_reason": {"type": "string", "minLength": 1, "maxLength": 2000}, "blocked_question": {"type": "string", "minLength": 1, "maxLength": 1000}, "proposed_verbs": {"type": "array", "maxItems": 10, "items": {"type": "object", "additionalProperties": false, "required": ["kind", "title", "outcome", "why", "priority"], "properties": {"kind": {"type": "string", "pattern": "^verb\\.[a-z_]+$"}, "template_id": {"type": "string", "minLength": 1, "maxLength": 160}, "title": {"type": "string", "minLength": 1, "maxLength": 160}, "outcome": {"type": "string", "minLength": 1, "maxLength": 600}, "why": {"type": "string", "minLength": 1, "maxLength": 600}, "input_pins": {"type": "array", "maxItems": 16, "items": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}, "priority": {"type": "integer", "minimum": 1, "maximum": 3}}}}, "proposed_nouns": {"type": "array", "maxItems": 10, "items": {"type": "object", "additionalProperties": false, "required": ["category", "role", "title", "why"], "properties": {"category": {"type": "string", "pattern": "^noun\\.[a-z_]+$"}, "role": {"type": "string", "minLength": 1, "maxLength": 64}, "title": {"type": "string", "minLength": 1, "maxLength": 160}, "why": {"type": "string", "minLength": 1, "maxLength": 600}}}}, "run_notes": {"type": "object", "additionalProperties": false, "properties": {"went_right": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}, "went_wrong": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}, "next_steps": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}, "ideas": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}}}}}, "permitted_outcomes": ["run_scored", "no_change", "unknown", "blocked"], "outcome_meanings": {"run_scored": "The declared nouns were produced and every deterministic acceptance check passed.", "no_change": "The evidence does not justify producing or changing anything. Say why.", "unknown": "The question could not be answered with the permitted inputs. List what is missing and what would resolve it.", "blocked": "A required input, permission or capability is missing. Ask one precise question."}, "outcome_requirements": {"run_scored": ["artifacts"], "no_change": ["no_change_reason"], "unknown": ["unknowns"], "blocked": ["blocked_question"]}, "acceptance_checks": [{"id": "worked-example-reproduced", "check": "The scorer reproduces example.worked's expected metrics and marks exactly.", "method": "deterministic"}, {"id": "scores-repeatable", "check": "Scoring the same run twice gives identical output.", "method": "deterministic"}, {"id": "labels-verified", "check": "Every label digest matched the manifest.", "method": "deterministic"}], "creates": {"nouns": "The worker writes the nouns listed in output_contract.produces and returns their pins.", "verbs": "The worker proposes follow-up verbs in output.proposed_verbs, such as a breakdown into smaller verbs or a research follow-up, with the pins they need. Proposals are never enqueued by the worker; an authorized requester enqueues the ones worth doing."}, "guidance": {"required": ["Reproduce the worked example's expected results first; stop with blocked if they differ.", "Verify the label digests against the manifest before scoring.", "Score exactly as the procedure says: null or an error is wrong; cost per case is null if any cost is null.", "Return follow-up work the results justify as proposed_verbs and nouns worth creating later as proposed_nouns. Never enqueue them yourself; an authorized requester decides which to run."], "preferred": [], "suggested": [], "discovery_requested": false}, "required_capabilities": ["read permitted noun versions", "write candidate nouns in the requester's scope"], "bounds": {"max_attempts": 2, "max_elapsed_seconds": 600}, "execution": {"reasoning": "deterministic_tools", "model_use": "none", "inference_allowance": "Existing authorized worker allowance only. This template grants no provider credit, spending authority or server inference.", "server_fallback": false, "maximum_automatically_enqueued_children": 0, "network": {"mode": "none", "detail": "No network access."}, "external_effects": "none"}, "review": {"required": "owning_policy", "self_approval_permitted": false, "publication_permitted": false, "admission_permitted": false}, "instantiation": {"status": "template", "fill_before_enqueue": ["run_record", "manifest", "pass_marks"], "rule": "Replace each slot with exact NounRef pins, set the params, then recompute the idempotency key with scripts/build_packs.py --instantiate. Enqueue only with your own grants in a queue your owner assigned."}, "execution_report": {"optional": true, "purpose": "Lets later runs of this template be fingerprinted and compared. It is worker reported, not proof.", "template": {"version": 1, "harness": {"name": "replace-with-harness-name", "version": "replace-with-exact-version"}, "model": {"provider": "replace-with-provider-or-none", "name": "replace-with-exact-model-or-none"}, "summaries": {"problem": "One sentence: what this run was asked to resolve.", "input": "One sentence: which pinned inputs were read.", "solution": "One sentence: what was produced or why the outcome was no_change, unknown or blocked."}, "contract": {"input_schema": {"type": "object", "additionalProperties": false, "required": ["pinned", "slots", "params"], "properties": {"pinned": {"type": "array", "minItems": 3, "maxItems": 3, "items": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}, "slots": {"type": "object", "additionalProperties": false, "required": ["run_record", "manifest", "pass_marks"], "properties": {"run_record": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}, "manifest": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}, "pass_marks": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}}, "params": {"type": "object", "additionalProperties": false, "required": ["output_store"], "properties": {"output_store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}}}}}, "output_schema": {"type": "object", "additionalProperties": false, "required": ["outcome", "summary", "artifacts", "checks", "unknowns"], "properties": {"outcome": {"type": "string", "enum": ["run_scored", "no_change", "unknown", "blocked"]}, "summary": {"type": "string", "minLength": 1, "maxLength": 1200}, "artifacts": {"type": "array", "maxItems": 1, "items": {"type": "object", "additionalProperties": false, "required": ["role", "ref"], "properties": {"role": {"type": "string", "minLength": 1, "maxLength": 64}, "ref": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}, "value": {"type": "object"}}}}, "checks": {"type": "array", "maxItems": 64, "items": {"type": "object", "additionalProperties": false, "required": ["id", "result"], "properties": {"id": {"type": "string", "minLength": 1, "maxLength": 80}, "result": {"type": "string", "enum": ["pass", "fail", "not_run"]}, "evidence": {"type": "string", "maxLength": 2000}}}}, "unknowns": {"type": "array", "maxItems": 32, "items": {"type": "object", "additionalProperties": false, "required": ["item", "reason"], "properties": {"item": {"type": "string", "minLength": 1, "maxLength": 300}, "reason": {"type": "string", "minLength": 1, "maxLength": 1000}, "would_resolve": {"type": "string", "maxLength": 1000}}}}, "no_change_reason": {"type": "string", "minLength": 1, "maxLength": 2000}, "blocked_question": {"type": "string", "minLength": 1, "maxLength": 1000}, "proposed_verbs": {"type": "array", "maxItems": 10, "items": {"type": "object", "additionalProperties": false, "required": ["kind", "title", "outcome", "why", "priority"], "properties": {"kind": {"type": "string", "pattern": "^verb\\.[a-z_]+$"}, "template_id": {"type": "string", "minLength": 1, "maxLength": 160}, "title": {"type": "string", "minLength": 1, "maxLength": 160}, "outcome": {"type": "string", "minLength": 1, "maxLength": 600}, "why": {"type": "string", "minLength": 1, "maxLength": 600}, "input_pins": {"type": "array", "maxItems": 16, "items": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}, "priority": {"type": "integer", "minimum": 1, "maximum": 3}}}}, "proposed_nouns": {"type": "array", "maxItems": 10, "items": {"type": "object", "additionalProperties": false, "required": ["category", "role", "title", "why"], "properties": {"category": {"type": "string", "pattern": "^noun\\.[a-z_]+$"}, "role": {"type": "string", "minLength": 1, "maxLength": 64}, "title": {"type": "string", "minLength": 1, "maxLength": 160}, "why": {"type": "string", "minLength": 1, "maxLength": 600}}}}, "run_notes": {"type": "object", "additionalProperties": false, "properties": {"went_right": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}, "went_wrong": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}, "next_steps": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}, "ideas": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}}}}}}, "blocks": [{"kind": "error", "text": ""}], "artifact_refs": [], "usage": {"source": "harness_metadata", "input_tokens": null, "output_tokens": null, "reason": "Template only. Replace with measured counters or keep null with a reason."}}, "fill_rules": ["Copy contract.input_schema and contract.output_schema unchanged, so every run of this template shares one shape fingerprint.", "Add context_hash, environment_hash and toolchain_hash as lowercase SHA-256 digests when known; all five contract fields together allow exact matching.", "Name the exact harness version, provider and model. Deterministic or human work records the model name as none.", "Put what went wrong in an error block and in output.run_notes.went_wrong. The version-1 report has no fields for what went right, next steps or ideas, so those go in output.run_notes.", "Leave token counts null with a reason unless the harness or provider measured them."], "run_notes_field": "output.run_notes", "shape_sha256": "ca3c62eb527015103fb7744f6ebb787432fb3a43be1ae9b04a6a45c3cb06396a"}, "dimensions": {"determinism": 0.0, "evidence_grounding": 0.0, "token_efficiency": 0.0, "safety_compliance": 0.0, "verification_tier": "unverified"}, "dimension_basis": {"measured": false, "method": "unmeasured_floor", "reason": "Tessera requires these axes, so each is set to its lowest value and the tier is unverified. No quality is claimed. Replace them only with measured values and the evidence that produced them."}, "relations": [{"relation": "requires_noun", "target_id": "seedpack.evaluation-design/evaluation-design.v0.1.0.scoring.exact-match", "target_category": "noun.reflex_procedure", "note": ""}, {"relation": "requires_noun", "target_id": "seedpack.evaluation-design/evaluation-design.v0.1.0.method", "target_category": "noun.skill_pack", "note": ""}, {"relation": "requires_noun", "target_id": "seedpack.evaluation-design/evaluation-design.v0.1.0.example.worked", "target_category": "noun.benchmark_suite", "note": ""}, {"relation": "produces_noun", "target_id": "seedpack.evaluation-design/evaluation-design.v0.1.0.template.run-record", "target_category": "noun.eval_run", "note": "output scored_run"}], "follows_templates": ["evaluation-design.run-arm"]}},
{"queue": "conformance-seed-packs", "kind": "verb.solve_draft", "effect_class": "write-internal", "priority": 50, "idempotency_key": "seedpack:evaluation-design:report-evaluation:0673f728a22bcb4bdc01a9bae0a0ca61bff1f2f10c12aed4af9e56dbdec879dd", "depends_on": [], "producer_surface": "surface:tessera", "visibility_scope": "surface_only", "allowed_surfaces": [], "target_cohorts": ["all"], "tags": ["category:verb.solve_draft", "pack:evaluation-design", "template:report-evaluation", "claimant:agent", "status:quarantined"], "brief": {"category_id": "verb.solve_draft", "contract": "ta-amps.seed-verb-template.v1", "pack": "evaluation-design", "pack_version": "0.1.0", "template_id": "evaluation-design.report-evaluation", "title": "Write the evaluation report", "plain_summary": "Report every committed mark as met, missed or unknown, with per-arm metrics, stability, cost per case and a diagnosis for each missed mark.", "outcome": "An evaluation report; missed marks stay missed.", "claimant": "agent", "inputs": {"pinned": [{"store": "seedpack.evaluation-design", "key": "evaluation-design.v0.1.0.method", "version": 1, "value_sha256": "436329f710d068a71b3e6f7252a8f4ff875bda0d6eac92424526cb5712be11db"}], "slots": {"plan": null, "pass_marks": null, "scored_runs": []}, "params": {"output_store": "seedpack.evaluation-design.outputs"}}, "slot_specs": {"plan": {"category": "noun.benchmark_suite", "required": true, "many": false, "description": "Pin of the plan."}, "pass_marks": {"category": "noun.seed_hypothesis", "required": true, "many": false, "description": "Pin of the committed marks."}, "scored_runs": {"category": "noun.eval_run", "required": true, "many": true, "description": "Pins of every scored run."}}, "param_specs": {"output_store": {"description": "Store for the produced nouns. The default is a placeholder; use the store your owner assigned.", "schema": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "required": true, "default": "seedpack.evaluation-design.outputs"}}, "input_schema": {"type": "object", "additionalProperties": false, "required": ["pinned", "slots", "params"], "properties": {"pinned": {"type": "array", "minItems": 1, "maxItems": 1, "items": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}, "slots": {"type": "object", "additionalProperties": false, "required": ["plan", "pass_marks", "scored_runs"], "properties": {"plan": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}, "pass_marks": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}, "scored_runs": {"type": "array", "minItems": 1, "maxItems": 20, "items": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}}}, "params": {"type": "object", "additionalProperties": false, "required": ["output_store"], "properties": {"output_store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}}}}}, "output_contract": {"visibility_scope": "surface_only", "allowed_surfaces": [], "write_mode": "create_only", "store_param": "output_store", "key_rule": "evaluation-design.<role>.<task_id>[.<n>]", "expected_version": 0, "produces": [{"name": "report", "template": {"store": "seedpack.evaluation-design", "key": "evaluation-design.v0.1.0.template.evaluation-report", "version": 1, "value_sha256": "ac06027569cdd0cc1249ebdb3ed3b471c72bf03a08cf123027061cbf1c726096"}, "max": 1}], "delivery": "Write produced nouns surface_only under your own authenticated surface, which is the requester's assigned surface, so the requester and that surface's reviewers can read them. Never use public_commons or allowlist other surfaces. Return every produced noun's pin in output.artifacts; add its value when the requester asks for inline delivery."}, "output_schema": {"type": "object", "additionalProperties": false, "required": ["outcome", "summary", "artifacts", "checks", "unknowns"], "properties": {"outcome": {"type": "string", "enum": ["report_ready", "no_change", "unknown", "blocked"]}, "summary": {"type": "string", "minLength": 1, "maxLength": 1200}, "artifacts": {"type": "array", "maxItems": 1, "items": {"type": "object", "additionalProperties": false, "required": ["role", "ref"], "properties": {"role": {"type": "string", "minLength": 1, "maxLength": 64}, "ref": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}, "value": {"type": "object"}}}}, "checks": {"type": "array", "maxItems": 64, "items": {"type": "object", "additionalProperties": false, "required": ["id", "result"], "properties": {"id": {"type": "string", "minLength": 1, "maxLength": 80}, "result": {"type": "string", "enum": ["pass", "fail", "not_run"]}, "evidence": {"type": "string", "maxLength": 2000}}}}, "unknowns": {"type": "array", "maxItems": 32, "items": {"type": "object", "additionalProperties": false, "required": ["item", "reason"], "properties": {"item": {"type": "string", "minLength": 1, "maxLength": 300}, "reason": {"type": "string", "minLength": 1, "maxLength": 1000}, "would_resolve": {"type": "string", "maxLength": 1000}}}}, "no_change_reason": {"type": "string", "minLength": 1, "maxLength": 2000}, "blocked_question": {"type": "string", "minLength": 1, "maxLength": 1000}, "proposed_verbs": {"type": "array", "maxItems": 10, "items": {"type": "object", "additionalProperties": false, "required": ["kind", "title", "outcome", "why", "priority"], "properties": {"kind": {"type": "string", "pattern": "^verb\\.[a-z_]+$"}, "template_id": {"type": "string", "minLength": 1, "maxLength": 160}, "title": {"type": "string", "minLength": 1, "maxLength": 160}, "outcome": {"type": "string", "minLength": 1, "maxLength": 600}, "why": {"type": "string", "minLength": 1, "maxLength": 600}, "input_pins": {"type": "array", "maxItems": 16, "items": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}, "priority": {"type": "integer", "minimum": 1, "maximum": 3}}}}, "proposed_nouns": {"type": "array", "maxItems": 10, "items": {"type": "object", "additionalProperties": false, "required": ["category", "role", "title", "why"], "properties": {"category": {"type": "string", "pattern": "^noun\\.[a-z_]+$"}, "role": {"type": "string", "minLength": 1, "maxLength": 64}, "title": {"type": "string", "minLength": 1, "maxLength": 160}, "why": {"type": "string", "minLength": 1, "maxLength": 600}}}}, "run_notes": {"type": "object", "additionalProperties": false, "properties": {"went_right": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}, "went_wrong": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}, "next_steps": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}, "ideas": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}}}}}, "permitted_outcomes": ["report_ready", "no_change", "unknown", "blocked"], "outcome_meanings": {"report_ready": "The declared nouns were produced and every deterministic acceptance check passed.", "no_change": "The evidence does not justify producing or changing anything. Say why.", "unknown": "The question could not be answered with the permitted inputs. List what is missing and what would resolve it.", "blocked": "A required input, permission or capability is missing. Ask one precise question."}, "outcome_requirements": {"report_ready": ["artifacts"], "no_change": ["no_change_reason"], "unknown": ["unknowns"], "blocked": ["blocked_question"]}, "acceptance_checks": [{"id": "report-matches-template", "check": "The report validates against template.evaluation-report's schema.", "method": "deterministic"}, {"id": "all-marks", "check": "Every committed mark appears exactly once.", "method": "deterministic"}, {"id": "metrics-copied", "check": "Every metric equals the value in its scored run.", "method": "deterministic"}, {"id": "diagnoses-useful", "check": "Each diagnosis points at a likely cause someone could test.", "method": "judgment"}], "creates": {"nouns": "The worker writes the nouns listed in output_contract.produces and returns their pins.", "verbs": "The worker proposes follow-up verbs in output.proposed_verbs, such as a breakdown into smaller verbs or a research follow-up, with the pins they need. Proposals are never enqueued by the worker; an authorized requester enqueues the ones worth doing."}, "guidance": {"required": ["Verify the pins and that every arm has runs_per_arm scored runs; otherwise report the gap as unknown.", "Copy metrics from the scored runs; do not recompute them differently.", "Mark an arm unstable when its spread exceeds the plan's limit.", "Give each missed mark a short diagnosis. Never reword a missed mark as a partial pass.", "Return follow-up work the results justify as proposed_verbs and nouns worth creating later as proposed_nouns. Never enqueue them yourself; an authorized requester decides which to run."], "preferred": [], "suggested": [], "discovery_requested": false}, "required_capabilities": ["read permitted noun versions", "write candidate nouns in the requester's scope"], "bounds": {"max_attempts": 2, "max_elapsed_seconds": 1200}, "execution": {"reasoning": "authorized_worker", "model_use": "worker_allowance", "inference_allowance": "Existing authorized worker allowance only. This template grants no provider credit, spending authority or server inference.", "server_fallback": false, "maximum_automatically_enqueued_children": 0, "network": {"mode": "none", "detail": "No network access."}, "external_effects": "none"}, "review": {"required": "owning_policy", "self_approval_permitted": false, "publication_permitted": false, "admission_permitted": false}, "instantiation": {"status": "template", "fill_before_enqueue": ["plan", "pass_marks", "scored_runs"], "rule": "Replace each slot with exact NounRef pins, set the params, then recompute the idempotency key with scripts/build_packs.py --instantiate. Enqueue only with your own grants in a queue your owner assigned."}, "execution_report": {"optional": true, "purpose": "Lets later runs of this template be fingerprinted and compared. It is worker reported, not proof.", "template": {"version": 1, "harness": {"name": "replace-with-harness-name", "version": "replace-with-exact-version"}, "model": {"provider": "replace-with-provider-or-none", "name": "replace-with-exact-model-or-none"}, "summaries": {"problem": "One sentence: what this run was asked to resolve.", "input": "One sentence: which pinned inputs were read.", "solution": "One sentence: what was produced or why the outcome was no_change, unknown or blocked."}, "contract": {"input_schema": {"type": "object", "additionalProperties": false, "required": ["pinned", "slots", "params"], "properties": {"pinned": {"type": "array", "minItems": 1, "maxItems": 1, "items": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}, "slots": {"type": "object", "additionalProperties": false, "required": ["plan", "pass_marks", "scored_runs"], "properties": {"plan": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}, "pass_marks": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}, "scored_runs": {"type": "array", "minItems": 1, "maxItems": 20, "items": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}}}, "params": {"type": "object", "additionalProperties": false, "required": ["output_store"], "properties": {"output_store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}}}}}, "output_schema": {"type": "object", "additionalProperties": false, "required": ["outcome", "summary", "artifacts", "checks", "unknowns"], "properties": {"outcome": {"type": "string", "enum": ["report_ready", "no_change", "unknown", "blocked"]}, "summary": {"type": "string", "minLength": 1, "maxLength": 1200}, "artifacts": {"type": "array", "maxItems": 1, "items": {"type": "object", "additionalProperties": false, "required": ["role", "ref"], "properties": {"role": {"type": "string", "minLength": 1, "maxLength": 64}, "ref": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}, "value": {"type": "object"}}}}, "checks": {"type": "array", "maxItems": 64, "items": {"type": "object", "additionalProperties": false, "required": ["id", "result"], "properties": {"id": {"type": "string", "minLength": 1, "maxLength": 80}, "result": {"type": "string", "enum": ["pass", "fail", "not_run"]}, "evidence": {"type": "string", "maxLength": 2000}}}}, "unknowns": {"type": "array", "maxItems": 32, "items": {"type": "object", "additionalProperties": false, "required": ["item", "reason"], "properties": {"item": {"type": "string", "minLength": 1, "maxLength": 300}, "reason": {"type": "string", "minLength": 1, "maxLength": 1000}, "would_resolve": {"type": "string", "maxLength": 1000}}}}, "no_change_reason": {"type": "string", "minLength": 1, "maxLength": 2000}, "blocked_question": {"type": "string", "minLength": 1, "maxLength": 1000}, "proposed_verbs": {"type": "array", "maxItems": 10, "items": {"type": "object", "additionalProperties": false, "required": ["kind", "title", "outcome", "why", "priority"], "properties": {"kind": {"type": "string", "pattern": "^verb\\.[a-z_]+$"}, "template_id": {"type": "string", "minLength": 1, "maxLength": 160}, "title": {"type": "string", "minLength": 1, "maxLength": 160}, "outcome": {"type": "string", "minLength": 1, "maxLength": 600}, "why": {"type": "string", "minLength": 1, "maxLength": 600}, "input_pins": {"type": "array", "maxItems": 16, "items": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}, "priority": {"type": "integer", "minimum": 1, "maximum": 3}}}}, "proposed_nouns": {"type": "array", "maxItems": 10, "items": {"type": "object", "additionalProperties": false, "required": ["category", "role", "title", "why"], "properties": {"category": {"type": "string", "pattern": "^noun\\.[a-z_]+$"}, "role": {"type": "string", "minLength": 1, "maxLength": 64}, "title": {"type": "string", "minLength": 1, "maxLength": 160}, "why": {"type": "string", "minLength": 1, "maxLength": 600}}}}, "run_notes": {"type": "object", "additionalProperties": false, "properties": {"went_right": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}, "went_wrong": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}, "next_steps": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}, "ideas": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}}}}}}, "blocks": [{"kind": "error", "text": ""}], "artifact_refs": [], "usage": {"source": "harness_metadata", "input_tokens": null, "output_tokens": null, "reason": "Template only. Replace with measured counters or keep null with a reason."}}, "fill_rules": ["Copy contract.input_schema and contract.output_schema unchanged, so every run of this template shares one shape fingerprint.", "Add context_hash, environment_hash and toolchain_hash as lowercase SHA-256 digests when known; all five contract fields together allow exact matching.", "Name the exact harness version, provider and model. Deterministic or human work records the model name as none.", "Put what went wrong in an error block and in output.run_notes.went_wrong. The version-1 report has no fields for what went right, next steps or ideas, so those go in output.run_notes.", "Leave token counts null with a reason unless the harness or provider measured them."], "run_notes_field": "output.run_notes", "shape_sha256": "562561c14b0afa9ac44430873ca0f1ee3700866e2345d23ec54160e00e96c957"}, "dimensions": {"determinism": 0.0, "evidence_grounding": 0.0, "token_efficiency": 0.0, "safety_compliance": 0.0, "verification_tier": "unverified"}, "dimension_basis": {"measured": false, "method": "unmeasured_floor", "reason": "Tessera requires these axes, so each is set to its lowest value and the tier is unverified. No quality is claimed. Replace them only with measured values and the evidence that produced them."}, "relations": [{"relation": "requires_noun", "target_id": "seedpack.evaluation-design/evaluation-design.v0.1.0.method", "target_category": "noun.skill_pack", "note": ""}, {"relation": "produces_noun", "target_id": "seedpack.evaluation-design/evaluation-design.v0.1.0.template.evaluation-report", "target_category": "noun.eval_run", "note": "output report"}], "follows_templates": ["evaluation-design.score-run"]}},
{"queue": "conformance-seed-packs", "kind": "verb.verify_gate", "effect_class": "write-internal", "priority": 50, "idempotency_key": "seedpack:evaluation-design:review-evaluation:cdac1bf1f6e5f61c46079855b411e20b6e3053647996090e86806548312bc436", "depends_on": [], "producer_surface": "surface:tessera", "visibility_scope": "surface_only", "allowed_surfaces": [], "target_cohorts": ["all"], "tags": ["category:verb.verify_gate", "pack:evaluation-design", "template:review-evaluation", "claimant:agent", "status:quarantined"], "brief": {"category_id": "verb.verify_gate", "contract": "ta-amps.seed-verb-template.v1", "pack": "evaluation-design", "pack_version": "0.1.0", "template_id": "evaluation-design.review-evaluation", "title": "Review the evaluation", "plain_summary": "A different worker rechecks the report against the scored runs, the committed marks and the method's rules.", "outcome": "Review notes, with any rule the evaluation broke.", "claimant": "agent", "inputs": {"pinned": [{"store": "seedpack.evaluation-design", "key": "evaluation-design.v0.1.0.method", "version": 1, "value_sha256": "436329f710d068a71b3e6f7252a8f4ff875bda0d6eac92424526cb5712be11db"}, {"store": "seedpack.evaluation-design", "key": "evaluation-design.v0.1.0.scoring.exact-match", "version": 1, "value_sha256": "677024a78f60eef2c8391d61e0ca0a77e71a3443477b6bd6d64d0d400a68ae27"}], "slots": {"report": null, "scored_runs": [], "pass_marks": null}, "params": {"output_store": "seedpack.evaluation-design.outputs"}}, "slot_specs": {"report": {"category": "noun.eval_run", "required": true, "many": false, "description": "Pin of the evaluation report."}, "scored_runs": {"category": "noun.eval_run", "required": true, "many": true, "description": "Pins of the scored runs."}, "pass_marks": {"category": "noun.seed_hypothesis", "required": true, "many": false, "description": "Pin of the committed marks."}}, "param_specs": {"output_store": {"description": "Store for the produced nouns. The default is a placeholder; use the store your owner assigned.", "schema": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "required": true, "default": "seedpack.evaluation-design.outputs"}}, "input_schema": {"type": "object", "additionalProperties": false, "required": ["pinned", "slots", "params"], "properties": {"pinned": {"type": "array", "minItems": 2, "maxItems": 2, "items": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}, "slots": {"type": "object", "additionalProperties": false, "required": ["report", "scored_runs", "pass_marks"], "properties": {"report": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}, "scored_runs": {"type": "array", "minItems": 1, "maxItems": 20, "items": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}, "pass_marks": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}}, "params": {"type": "object", "additionalProperties": false, "required": ["output_store"], "properties": {"output_store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}}}}}, "output_contract": {"visibility_scope": "surface_only", "allowed_surfaces": [], "write_mode": "create_only", "store_param": "output_store", "key_rule": "evaluation-design.<role>.<task_id>[.<n>]", "expected_version": 0, "produces": [{"name": "review", "template": {"store": "seedpack.evaluation-design", "key": "evaluation-design.v0.1.0.template.evaluation-report", "version": 1, "value_sha256": "ac06027569cdd0cc1249ebdb3ed3b471c72bf03a08cf123027061cbf1c726096"}, "max": 1}], "delivery": "Write produced nouns surface_only under your own authenticated surface, which is the requester's assigned surface, so the requester and that surface's reviewers can read them. Never use public_commons or allowlist other surfaces. Return every produced noun's pin in output.artifacts; add its value when the requester asks for inline delivery."}, "output_schema": {"type": "object", "additionalProperties": false, "required": ["outcome", "summary", "artifacts", "checks", "unknowns"], "properties": {"outcome": {"type": "string", "enum": ["review_ready", "no_change", "unknown", "blocked"]}, "summary": {"type": "string", "minLength": 1, "maxLength": 1200}, "artifacts": {"type": "array", "maxItems": 1, "items": {"type": "object", "additionalProperties": false, "required": ["role", "ref"], "properties": {"role": {"type": "string", "minLength": 1, "maxLength": 64}, "ref": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}, "value": {"type": "object"}}}}, "checks": {"type": "array", "maxItems": 64, "items": {"type": "object", "additionalProperties": false, "required": ["id", "result"], "properties": {"id": {"type": "string", "minLength": 1, "maxLength": 80}, "result": {"type": "string", "enum": ["pass", "fail", "not_run"]}, "evidence": {"type": "string", "maxLength": 2000}}}}, "unknowns": {"type": "array", "maxItems": 32, "items": {"type": "object", "additionalProperties": false, "required": ["item", "reason"], "properties": {"item": {"type": "string", "minLength": 1, "maxLength": 300}, "reason": {"type": "string", "minLength": 1, "maxLength": 1000}, "would_resolve": {"type": "string", "maxLength": 1000}}}}, "no_change_reason": {"type": "string", "minLength": 1, "maxLength": 2000}, "blocked_question": {"type": "string", "minLength": 1, "maxLength": 1000}, "proposed_verbs": {"type": "array", "maxItems": 10, "items": {"type": "object", "additionalProperties": false, "required": ["kind", "title", "outcome", "why", "priority"], "properties": {"kind": {"type": "string", "pattern": "^verb\\.[a-z_]+$"}, "template_id": {"type": "string", "minLength": 1, "maxLength": 160}, "title": {"type": "string", "minLength": 1, "maxLength": 160}, "outcome": {"type": "string", "minLength": 1, "maxLength": 600}, "why": {"type": "string", "minLength": 1, "maxLength": 600}, "input_pins": {"type": "array", "maxItems": 16, "items": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}, "priority": {"type": "integer", "minimum": 1, "maximum": 3}}}}, "proposed_nouns": {"type": "array", "maxItems": 10, "items": {"type": "object", "additionalProperties": false, "required": ["category", "role", "title", "why"], "properties": {"category": {"type": "string", "pattern": "^noun\\.[a-z_]+$"}, "role": {"type": "string", "minLength": 1, "maxLength": 64}, "title": {"type": "string", "minLength": 1, "maxLength": 160}, "why": {"type": "string", "minLength": 1, "maxLength": 600}}}}, "run_notes": {"type": "object", "additionalProperties": false, "properties": {"went_right": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}, "went_wrong": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}, "next_steps": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}, "ideas": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}}}}}, "permitted_outcomes": ["review_ready", "no_change", "unknown", "blocked"], "outcome_meanings": {"review_ready": "The declared nouns were produced and every deterministic acceptance check passed.", "no_change": "The evidence does not justify producing or changing anything. Say why.", "unknown": "The question could not be answered with the permitted inputs. List what is missing and what would resolve it.", "blocked": "A required input, permission or capability is missing. Ask one precise question."}, "outcome_requirements": {"review_ready": ["artifacts"], "no_change": ["no_change_reason"], "unknown": ["unknowns"], "blocked": ["blocked_question"]}, "acceptance_checks": [{"id": "marks-recomputed", "check": "Every mark's result matches a fresh recomputation.", "method": "deterministic"}, {"id": "distinct-reviewer", "check": "The reviewing key differs from the report's and scorer's keys.", "method": "deterministic"}, {"id": "rules-checked", "check": "Each method rule is marked kept or broken with evidence.", "method": "judgment"}], "creates": {"nouns": "The worker writes the nouns listed in output_contract.produces and returns their pins.", "verbs": "The worker proposes follow-up verbs in output.proposed_verbs, such as a breakdown into smaller verbs or a research follow-up, with the pins they need. Proposals are never enqueued by the worker; an authorized requester enqueues the ones worth doing."}, "guidance": {"required": ["Confirm you did not write the report or score its runs; otherwise return blocked.", "Recompute each mark from the scored runs and compare.", "Check the method's rules: marks before runs, opaque ids, two runs per arm, nulls honest.", "Return follow-up work the results justify as proposed_verbs and nouns worth creating later as proposed_nouns. Never enqueue them yourself; an authorized requester decides which to run."], "preferred": [], "suggested": [], "discovery_requested": false}, "required_capabilities": ["read permitted noun versions", "write candidate nouns in the requester's scope"], "bounds": {"max_attempts": 2, "max_elapsed_seconds": 1200}, "execution": {"reasoning": "authorized_worker", "model_use": "worker_allowance", "inference_allowance": "Existing authorized worker allowance only. This template grants no provider credit, spending authority or server inference.", "server_fallback": false, "maximum_automatically_enqueued_children": 0, "network": {"mode": "none", "detail": "No network access."}, "external_effects": "none"}, "review": {"required": "distinct_reviewer", "self_approval_permitted": false, "publication_permitted": false, "admission_permitted": false}, "instantiation": {"status": "template", "fill_before_enqueue": ["report", "scored_runs", "pass_marks"], "rule": "Replace each slot with exact NounRef pins, set the params, then recompute the idempotency key with scripts/build_packs.py --instantiate. Enqueue only with your own grants in a queue your owner assigned."}, "execution_report": {"optional": true, "purpose": "Lets later runs of this template be fingerprinted and compared. It is worker reported, not proof.", "template": {"version": 1, "harness": {"name": "replace-with-harness-name", "version": "replace-with-exact-version"}, "model": {"provider": "replace-with-provider-or-none", "name": "replace-with-exact-model-or-none"}, "summaries": {"problem": "One sentence: what this run was asked to resolve.", "input": "One sentence: which pinned inputs were read.", "solution": "One sentence: what was produced or why the outcome was no_change, unknown or blocked."}, "contract": {"input_schema": {"type": "object", "additionalProperties": false, "required": ["pinned", "slots", "params"], "properties": {"pinned": {"type": "array", "minItems": 2, "maxItems": 2, "items": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}, "slots": {"type": "object", "additionalProperties": false, "required": ["report", "scored_runs", "pass_marks"], "properties": {"report": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}, "scored_runs": {"type": "array", "minItems": 1, "maxItems": 20, "items": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}, "pass_marks": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}}, "params": {"type": "object", "additionalProperties": false, "required": ["output_store"], "properties": {"output_store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}}}}}, "output_schema": {"type": "object", "additionalProperties": false, "required": ["outcome", "summary", "artifacts", "checks", "unknowns"], "properties": {"outcome": {"type": "string", "enum": ["review_ready", "no_change", "unknown", "blocked"]}, "summary": {"type": "string", "minLength": 1, "maxLength": 1200}, "artifacts": {"type": "array", "maxItems": 1, "items": {"type": "object", "additionalProperties": false, "required": ["role", "ref"], "properties": {"role": {"type": "string", "minLength": 1, "maxLength": 64}, "ref": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}, "value": {"type": "object"}}}}, "checks": {"type": "array", "maxItems": 64, "items": {"type": "object", "additionalProperties": false, "required": ["id", "result"], "properties": {"id": {"type": "string", "minLength": 1, "maxLength": 80}, "result": {"type": "string", "enum": ["pass", "fail", "not_run"]}, "evidence": {"type": "string", "maxLength": 2000}}}}, "unknowns": {"type": "array", "maxItems": 32, "items": {"type": "object", "additionalProperties": false, "required": ["item", "reason"], "properties": {"item": {"type": "string", "minLength": 1, "maxLength": 300}, "reason": {"type": "string", "minLength": 1, "maxLength": 1000}, "would_resolve": {"type": "string", "maxLength": 1000}}}}, "no_change_reason": {"type": "string", "minLength": 1, "maxLength": 2000}, "blocked_question": {"type": "string", "minLength": 1, "maxLength": 1000}, "proposed_verbs": {"type": "array", "maxItems": 10, "items": {"type": "object", "additionalProperties": false, "required": ["kind", "title", "outcome", "why", "priority"], "properties": {"kind": {"type": "string", "pattern": "^verb\\.[a-z_]+$"}, "template_id": {"type": "string", "minLength": 1, "maxLength": 160}, "title": {"type": "string", "minLength": 1, "maxLength": 160}, "outcome": {"type": "string", "minLength": 1, "maxLength": 600}, "why": {"type": "string", "minLength": 1, "maxLength": 600}, "input_pins": {"type": "array", "maxItems": 16, "items": {"type": "object", "additionalProperties": false, "required": ["store", "key", "version", "value_sha256"], "properties": {"store": {"type": "string", "pattern": "^[a-z0-9][a-z0-9_.-]{0,63}$"}, "key": {"type": "string", "pattern": "^[A-Za-z0-9][A-Za-z0-9_.:-]{0,255}$"}, "version": {"type": "integer", "minimum": 1}, "value_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}}, "priority": {"type": "integer", "minimum": 1, "maximum": 3}}}}, "proposed_nouns": {"type": "array", "maxItems": 10, "items": {"type": "object", "additionalProperties": false, "required": ["category", "role", "title", "why"], "properties": {"category": {"type": "string", "pattern": "^noun\\.[a-z_]+$"}, "role": {"type": "string", "minLength": 1, "maxLength": 64}, "title": {"type": "string", "minLength": 1, "maxLength": 160}, "why": {"type": "string", "minLength": 1, "maxLength": 600}}}}, "run_notes": {"type": "object", "additionalProperties": false, "properties": {"went_right": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}, "went_wrong": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}, "next_steps": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}, "ideas": {"type": "array", "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 500}}}}}}}, "blocks": [{"kind": "error", "text": ""}], "artifact_refs": [], "usage": {"source": "harness_metadata", "input_tokens": null, "output_tokens": null, "reason": "Template only. Replace with measured counters or keep null with a reason."}}, "fill_rules": ["Copy contract.input_schema and contract.output_schema unchanged, so every run of this template shares one shape fingerprint.", "Add context_hash, environment_hash and toolchain_hash as lowercase SHA-256 digests when known; all five contract fields together allow exact matching.", "Name the exact harness version, provider and model. Deterministic or human work records the model name as none.", "Put what went wrong in an error block and in output.run_notes.went_wrong. The version-1 report has no fields for what went right, next steps or ideas, so those go in output.run_notes.", "Leave token counts null with a reason unless the harness or provider measured them."], "run_notes_field": "output.run_notes", "shape_sha256": "ebcc29470fa209f0096d5078fa3fd7576d9b5fb2087958a8464a37151d9c4164"}, "dimensions": {"determinism": 0.0, "evidence_grounding": 0.0, "token_efficiency": 0.0, "safety_compliance": 0.0, "verification_tier": "unverified"}, "dimension_basis": {"measured": false, "method": "unmeasured_floor", "reason": "Tessera requires these axes, so each is set to its lowest value and the tier is unverified. No quality is claimed. Replace them only with measured values and the evidence that produced them."}, "relations": [{"relation": "requires_noun", "target_id": "seedpack.evaluation-design/evaluation-design.v0.1.0.method", "target_category": "noun.skill_pack", "note": ""}, {"relation": "requires_noun", "target_id": "seedpack.evaluation-design/evaluation-design.v0.1.0.scoring.exact-match", "target_category": "noun.reflex_procedure", "note": ""}, {"relation": "produces_noun", "target_id": "seedpack.evaluation-design/evaluation-design.v0.1.0.template.evaluation-report", "target_category": "noun.eval_run", "note": "output review"}], "follows_templates": ["evaluation-design.report-evaluation"]}}
]}