{"format": "ta-amps.seed-pack.nouns.v1", "pack_id": "benchmark-scores", "status": "proposal_only", "records": [
{"store": "seedpack.benchmark-scores", "key": "benchmark-scores.v0.1.0.pack", "expected_version": 0, "producer_surface": "surface:tessera", "visibility_scope": "surface_only", "allowed_surfaces": [], "target_cohorts": ["all"], "tags": ["category:noun.skill_pack", "pack:benchmark-scores", "role:pack", "content:authored", "spdx:MIT", "status:quarantined"], "value": {"category_id": "noun.skill_pack", "role": "pack", "pack": "benchmark-scores", "pack_version": "0.1.0", "title": "A collection of nouns and verbs for keeping benchmark scores current and comparable", "lets_workers": "Keep a set of benchmark scores you care about current and honest: record every score with its source, date, hardware, setup and method, refresh them daily, flag the stale ones, and check important claims on your own hardware.", "flow": [{"verb": "benchmark-scores.collect-scores", "when": "When a watched benchmark or subject has a new reported score."}, {"verb": "benchmark-scores.refresh-scores", "when": "Daily, from the schedule proposal: re-read each watched score's source."}, {"verb": "benchmark-scores.flag-stale", "when": "Daily after refresh; deterministic, and it proposes refresh or comparison verbs."}, {"verb": "benchmark-scores.compare-on-profile", "when": "For claims that could change a decision; two runs on your hardware."}, {"verb": "benchmark-scores.situate-score", "when": "When a new score may agree or disagree with earlier ones."}], "schedule": {"store": "seedpack.benchmark-scores", "key": "benchmark-scores.v0.1.0.schedule.daily-refresh", "version": 1, "value_sha256": "720193abd3b11de09252abb0c049004e2be936076b435d2cc73ba9436df34f5a"}, "rule": "A score never stands alone: no source, date, setup and method means it cannot be compared.", "status": "proposal", "admission": "not_admitted", "content_kind": "authored_guidance", "license": "MIT", "provenance": {"source": "original ta-amps authored proposal", "license": "MIT", "generator": "scripts/build_packs.py", "independently_verified": false}, "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.benchmark-scores/benchmark-scores.v0.1.0.method", "target_category": "noun.skill_pack", "note": ""}, {"relation": "requires_noun", "target_id": "seedpack.benchmark-scores/benchmark-scores.v0.1.0.staleness-policy", "target_category": "noun.reflex_procedure", "note": ""}, {"relation": "requires_noun", "target_id": "seedpack.benchmark-scores/benchmark-scores.v0.1.0.example.staleness", "target_category": "noun.benchmark_suite", "note": ""}, {"relation": "requires_noun", "target_id": "seedpack.benchmark-scores/benchmark-scores.v0.1.0.schedule.daily-refresh", "target_category": "noun.seed_hypothesis", "note": ""}, {"relation": "requires_noun", "target_id": "seedpack.benchmark-scores/benchmark-scores.v0.1.0.template.score", "target_category": "noun.eval_run", "note": ""}, {"relation": "requires_noun", "target_id": "seedpack.benchmark-scores/benchmark-scores.v0.1.0.template.benchmark", "target_category": "noun.benchmark_suite", "note": ""}, {"relation": "requires_noun", "target_id": "seedpack.benchmark-scores/benchmark-scores.v0.1.0.template.watch-list", "target_category": "noun.seed_hypothesis", "note": ""}, {"relation": "requires_noun", "target_id": "seedpack.benchmark-scores/benchmark-scores.v0.1.0.template.comparison", "target_category": "noun.eval_run", "note": ""}, {"relation": "requires_noun", "target_id": "seedpack.benchmark-scores/benchmark-scores.v0.1.0.template.staleness-report", "target_category": "noun.eval_run", "note": ""}]}},
{"store": "seedpack.benchmark-scores", "key": "benchmark-scores.v0.1.0.method", "expected_version": 0, "producer_surface": "surface:tessera", "visibility_scope": "surface_only", "allowed_surfaces": [], "target_cohorts": ["all"], "tags": ["category:noun.skill_pack", "pack:benchmark-scores", "role:method", "content:authored", "spdx:MIT", "status:quarantined"], "value": {"category_id": "noun.skill_pack", "role": "method", "pack": "benchmark-scores", "pack_version": "0.1.0", "title": "Comparing benchmark scores honestly", "category_reason": "A reusable method with guardrails for every verb in this pack.", "rules": ["Two scores are comparable only with the same benchmark version, split and metric definition, and a known setup.", "Keep who reported a score: vendor, paper, leaderboard, community or our own run.", "Unknown setup fields stay null; a missing hardware field makes a score unverified, not wrong.", "Our own reproduction runs twice under committed pass marks, on the pinned profile's hardware.", "Never average unrelated leaderboards into one number.", "Read only public pages and permitted exports. Do not sign up, log in, submit results or contact anyone."], "reproduction_method": {"store": "seedpack.evaluation-design", "key": "evaluation-design.v0.1.0.method", "version": 1, "value_sha256": "436329f710d068a71b3e6f7252a8f4ff875bda0d6eac92424526cb5712be11db"}, "status": "proposal", "admission": "not_admitted", "content_kind": "authored_guidance", "license": "MIT", "provenance": {"source": "original ta-amps authored proposal", "license": "MIT", "generator": "scripts/build_packs.py", "independently_verified": false}, "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": ""}]}},
{"store": "seedpack.benchmark-scores", "key": "benchmark-scores.v0.1.0.staleness-policy", "expected_version": 0, "producer_surface": "surface:tessera", "visibility_scope": "surface_only", "allowed_surfaces": [], "target_cohorts": ["all"], "tags": ["category:noun.reflex_procedure", "pack:benchmark-scores", "role:staleness_policy", "content:authored", "spdx:MIT", "status:quarantined"], "value": {"category_id": "noun.reflex_procedure", "role": "staleness_policy", "pack": "benchmark-scores", "pack_version": "0.1.0", "title": "Staleness policy: when a benchmark score needs refreshing", "category_reason": "A zero-call deterministic rule: the same scores and date always give the same flags.", "max_age_days": {"vendor": 30, "leaderboard": 30, "community": 30, "paper": 90, "our_run": 60}, "contradiction_tolerance_ratio": 0.1, "flags": {"older_than_max_age": "The score was observed longer ago than its reporter's max age.", "subject_changed": "The subject now has a newer exact version than the one scored.", "benchmark_changed": "The benchmark now has a newer version than the one scored.", "contradicted_by_our_run": "Our own run differs from the claim by more than the tolerance ratio.", "missing_setup": "The hardware is unknown, so the score cannot be compared with ours."}, "status_rule": "contradicted if contradicted_by_our_run; otherwise stale if older_than_max_age, subject_changed or benchmark_changed; otherwise unverified if missing_setup; otherwise current.", "model_calls": 0, "status": "proposal", "admission": "not_admitted", "content_kind": "authored_guidance", "license": "MIT", "provenance": {"source": "original ta-amps authored proposal", "license": "MIT", "generator": "scripts/build_packs.py", "independently_verified": false}, "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": []}},
{"store": "seedpack.benchmark-scores", "key": "benchmark-scores.v0.1.0.example.staleness", "expected_version": 0, "producer_surface": "surface:tessera", "visibility_scope": "surface_only", "allowed_surfaces": [], "target_cohorts": ["all"], "tags": ["category:noun.benchmark_suite", "pack:benchmark-scores", "role:worked_example", "synthetic:true", "spdx:MIT", "status:quarantined"], "value": {"category_id": "noun.benchmark_suite", "role": "worked_example", "pack": "benchmark-scores", "pack_version": "0.1.0", "title": "SYNTHETIC worked example: six scores and their staleness flags", "category_reason": "A golden fixture with recorded expected flags that the validator recomputes.", "fixture_kind": "staleness_flags", "as_of": "2026-10-29", "policy": {"max_age_days": {"vendor": 30, "leaderboard": 30, "community": 30, "paper": 90, "our_run": 60}, "contradiction_tolerance_ratio": 0.1, "flags": {"older_than_max_age": "The score was observed longer ago than its reporter's max age.", "subject_changed": "The subject now has a newer exact version than the one scored.", "benchmark_changed": "The benchmark now has a newer version than the one scored.", "contradicted_by_our_run": "Our own run differs from the claim by more than the tolerance ratio.", "missing_setup": "The hardware is unknown, so the score cannot be compared with ours."}, "status_rule": "contradicted if contradicted_by_our_run; otherwise stale if older_than_max_age, subject_changed or benchmark_changed; otherwise unverified if missing_setup; otherwise current."}, "scores": [{"id": "s1", "reported_by": "leaderboard", "observed_at": "2026-10-20", "subject_version": "1.2", "current_subject_version": "1.2", "benchmark_version": "v2", "current_benchmark_version": "v2", "value": 71.0, "our_value": null, "hardware": "SYNTHETIC GPU A"}, {"id": "s2", "reported_by": "vendor", "observed_at": "2026-08-01", "subject_version": "3.1", "current_subject_version": "3.1", "benchmark_version": "v2", "current_benchmark_version": "v2", "value": 12.5, "our_value": null, "hardware": "SYNTHETIC GPU B"}, {"id": "s3", "reported_by": "paper", "observed_at": "2026-09-01", "subject_version": "1.0", "current_subject_version": "2.0", "benchmark_version": "v2", "current_benchmark_version": "v2", "value": 0.82, "our_value": null, "hardware": "SYNTHETIC cluster"}, {"id": "s4", "reported_by": "leaderboard", "observed_at": "2026-10-25", "subject_version": "5.0", "current_subject_version": "5.0", "benchmark_version": "v1", "current_benchmark_version": "v2", "value": 55.0, "our_value": null, "hardware": "SYNTHETIC GPU A"}, {"id": "s5", "reported_by": "vendor", "observed_at": "2026-10-28", "subject_version": "2.4", "current_subject_version": "2.4", "benchmark_version": "v2", "current_benchmark_version": "v2", "value": 80.0, "our_value": 60.0, "hardware": "SYNTHETIC GPU C"}, {"id": "s6", "reported_by": "community", "observed_at": "2026-10-01", "subject_version": "0.9", "current_subject_version": "0.9", "benchmark_version": "v2", "current_benchmark_version": "v2", "value": 33.0, "our_value": null, "hardware": null}], "expected": {"s1": {"flags": [], "status": "current"}, "s2": {"flags": ["older_than_max_age"], "status": "stale"}, "s3": {"flags": ["subject_changed"], "status": "stale"}, "s4": {"flags": ["benchmark_changed"], "status": "stale"}, "s5": {"flags": ["contradicted_by_our_run"], "status": "contradicted"}, "s6": {"flags": ["missing_setup"], "status": "unverified"}}, "reading": ["s2 is 89 days old against a vendor limit of 30.", "s5 differs from our run by 25 percent, beyond the 10 percent tolerance.", "s6 has no hardware, so it stays unverified rather than stale."], "status": "proposal", "admission": "not_admitted", "content_kind": "synthetic_example", "license": "MIT", "provenance": {"source": "original ta-amps authored proposal", "license": "MIT", "generator": "scripts/build_packs.py", "independently_verified": false}, "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.benchmark-scores/benchmark-scores.v0.1.0.staleness-policy", "target_category": "noun.reflex_procedure", "note": ""}]}},
{"store": "seedpack.benchmark-scores", "key": "benchmark-scores.v0.1.0.template.score", "expected_version": 0, "producer_surface": "surface:tessera", "visibility_scope": "surface_only", "allowed_surfaces": [], "target_cohorts": ["all"], "tags": ["category:noun.eval_run", "pack:benchmark-scores", "role:template", "content:authored", "spdx:MIT", "status:quarantined"], "value": {"category_id": "noun.eval_run", "role": "template", "pack": "benchmark-scores", "pack_version": "0.1.0", "title": "Template: one benchmark score with its source, date, hardware, setup and method", "category_reason": "A score is a null-honest scorecard entry with its provenance.", "describes": "one benchmark score with its source, date, hardware, setup and method", "schema": {"type": "object", "additionalProperties": false, "required": ["benchmark", "subject", "value", "unit", "reported_by", "source", "setup", "method", "status", "checked_at"], "properties": {"benchmark": {"type": "object", "additionalProperties": false, "required": ["name", "version", "metric", "direction"], "properties": {"name": {"type": "string", "minLength": 1, "maxLength": 160}, "version": {"type": "string", "minLength": 1, "maxLength": 60}, "split": {"type": "string", "minLength": 0, "maxLength": 60}, "metric": {"type": "string", "minLength": 1, "maxLength": 120}, "direction": {"type": "string", "enum": ["higher", "lower"]}}}, "subject": {"type": "object", "additionalProperties": false, "required": ["name", "exact_version", "kind"], "properties": {"name": {"type": "string", "minLength": 1, "maxLength": 160}, "exact_version": {"type": "string", "minLength": 1, "maxLength": 120}, "kind": {"type": "string", "enum": ["model", "tool", "harness", "library", "service"]}}}, "value": {"type": ["number", "null"]}, "unit": {"type": "string", "minLength": 1, "maxLength": 40}, "reported_by": {"type": "string", "enum": ["vendor", "paper", "leaderboard", "community", "our_run"]}, "source": {"type": "object", "additionalProperties": false, "required": ["ref", "retrieved_at"], "properties": {"ref": {"type": "string", "minLength": 1, "maxLength": 400}, "retrieved_at": {"type": "string", "pattern": "^[0-9]{4}-[0-9]{2}-[0-9]{2}", "maxLength": 40}, "published_at": {"type": "string", "minLength": 0, "maxLength": 40}, "content_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}}}, "setup": {"type": "object", "additionalProperties": false, "required": ["hardware"], "properties": {"hardware": {"type": ["string", "null"], "maxLength": 200}, "accelerator_memory_gb": {"type": ["number", "null"]}, "cpu": {"type": ["string", "null"], "maxLength": 120}, "ram_gb": {"type": ["number", "null"]}, "os": {"type": ["string", "null"], "maxLength": 120}, "precision": {"type": ["string", "null"], "maxLength": 40}, "batch_size": {"type": ["integer", "null"], "minimum": 1}, "harness": {"type": ["string", "null"], "maxLength": 160}, "settings": {"type": ["string", "null"], "maxLength": 400}}}, "method": {"type": "string", "minLength": 1, "maxLength": 600}, "sample_size": {"type": ["integer", "null"], "minimum": 1}, "status": {"type": "string", "enum": ["current", "stale", "contradicted", "unverified"]}, "checked_at": {"type": "string", "pattern": "^[0-9]{4}-[0-9]{2}-[0-9]{2}", "maxLength": 40}}}, "rules": ["A score without hardware is unverified.", "value is null when the source gives no number."], "example": {"benchmark": {"name": "SYNTHETIC title-card render", "version": "v2", "metric": "wall time for a 10-second 1080p render", "direction": "lower"}, "subject": {"name": "SYNTHETIC motion library", "exact_version": "1.2.0", "kind": "library"}, "value": 4.8, "unit": "s", "reported_by": "vendor", "source": {"ref": "SYNTHETIC release note", "retrieved_at": "2026-10-29"}, "setup": {"hardware": null, "settings": "defaults"}, "method": "Reported by the vendor; the measurement method was not published.", "sample_size": null, "status": "unverified", "checked_at": "2026-10-29"}, "status": "proposal", "admission": "not_admitted", "content_kind": "authored_guidance", "license": "MIT", "provenance": {"source": "original ta-amps authored proposal", "license": "MIT", "generator": "scripts/build_packs.py", "independently_verified": false}, "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": []}},
{"store": "seedpack.benchmark-scores", "key": "benchmark-scores.v0.1.0.template.benchmark", "expected_version": 0, "producer_surface": "surface:tessera", "visibility_scope": "surface_only", "allowed_surfaces": [], "target_cohorts": ["all"], "tags": ["category:noun.benchmark_suite", "pack:benchmark-scores", "role:template", "content:authored", "spdx:MIT", "status:quarantined"], "value": {"category_id": "noun.benchmark_suite", "role": "template", "pack": "benchmark-scores", "pack_version": "0.1.0", "title": "Template: a benchmark definition: task, version, metrics and the setup fields needed to compare scores", "category_reason": "A benchmark definition is the harness and rubric that scores refer to.", "describes": "a benchmark definition: task, version, metrics and the setup fields needed to compare scores", "schema": {"type": "object", "additionalProperties": false, "required": ["name", "version", "task", "metrics", "required_setup_fields", "source"], "properties": {"name": {"type": "string", "minLength": 1, "maxLength": 160}, "version": {"type": "string", "minLength": 1, "maxLength": 60}, "task": {"type": "string", "minLength": 1, "maxLength": 600}, "metrics": {"type": "array", "minItems": 1, "maxItems": 10, "items": {"type": "object", "additionalProperties": false, "required": ["name", "unit", "direction"], "properties": {"name": {"type": "string", "minLength": 1, "maxLength": 120}, "unit": {"type": "string", "minLength": 1, "maxLength": 40}, "direction": {"type": "string", "enum": ["higher", "lower"]}}}}, "splits": {"type": "array", "minItems": 0, "maxItems": 10, "items": {"type": "string", "minLength": 1, "maxLength": 60}}, "required_setup_fields": {"type": "array", "minItems": 1, "maxItems": 12, "items": {"type": "string", "minLength": 1, "maxLength": 60}}, "source": {"type": "string", "minLength": 1, "maxLength": 400}}}, "rules": ["A new benchmark version is a new noun; scores name the version they used."], "status": "proposal", "admission": "not_admitted", "content_kind": "authored_guidance", "license": "MIT", "provenance": {"source": "original ta-amps authored proposal", "license": "MIT", "generator": "scripts/build_packs.py", "independently_verified": false}, "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": []}},
{"store": "seedpack.benchmark-scores", "key": "benchmark-scores.v0.1.0.template.watch-list", "expected_version": 0, "producer_surface": "surface:tessera", "visibility_scope": "surface_only", "allowed_surfaces": [], "target_cohorts": ["all"], "tags": ["category:noun.seed_hypothesis", "pack:benchmark-scores", "role:template", "content:authored", "spdx:MIT", "status:quarantined"], "value": {"category_id": "noun.seed_hypothesis", "role": "template", "pack": "benchmark-scores", "pack_version": "0.1.0", "title": "Template: which benchmarks and subjects to watch, why, and how old a score may get", "category_reason": "A watch list is a falsifiable bet about which scores will matter for decisions.", "describes": "which benchmarks and subjects to watch, why, and how old a score may get", "schema": {"type": "object", "additionalProperties": false, "required": ["profile_ref", "items"], "properties": {"profile_ref": {"type": "string", "minLength": 1, "maxLength": 300}, "items": {"type": "array", "minItems": 1, "maxItems": 50, "items": {"type": "object", "additionalProperties": false, "required": ["benchmark", "subjects", "why", "max_age_days"], "properties": {"benchmark": {"type": "string", "minLength": 1, "maxLength": 160}, "subjects": {"type": "array", "minItems": 1, "maxItems": 20, "items": {"type": "string", "minLength": 1, "maxLength": 160}}, "why": {"type": "string", "minLength": 1, "maxLength": 300}, "max_age_days": {"type": "integer", "minimum": 1, "maximum": 365}, "sources": {"type": "array", "minItems": 0, "maxItems": 10, "items": {"type": "string", "minLength": 1, "maxLength": 400}}}}}}}, "rules": ["Every item says why it matters to the pinned profile."], "status": "proposal", "admission": "not_admitted", "content_kind": "authored_guidance", "license": "MIT", "provenance": {"source": "original ta-amps authored proposal", "license": "MIT", "generator": "scripts/build_packs.py", "independently_verified": false}, "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": []}},
{"store": "seedpack.benchmark-scores", "key": "benchmark-scores.v0.1.0.template.comparison", "expected_version": 0, "producer_surface": "surface:tessera", "visibility_scope": "surface_only", "allowed_surfaces": [], "target_cohorts": ["all"], "tags": ["category:noun.eval_run", "pack:benchmark-scores", "role:template", "content:authored", "spdx:MIT", "status:quarantined"], "value": {"category_id": "noun.eval_run", "role": "template", "pack": "benchmark-scores", "pack_version": "0.1.0", "title": "Template: a claim checked on our own hardware: two runs, comparability and a verdict", "category_reason": "A comparison is a measured run against a reported claim.", "describes": "a claim checked on our own hardware: two runs, comparability and a verdict", "schema": {"type": "object", "additionalProperties": false, "required": ["claim_ref", "profile_ref", "our_runs", "comparable", "difference_ratio", "verdict", "reasons"], "properties": {"claim_ref": {"type": "string", "minLength": 1, "maxLength": 300}, "profile_ref": {"type": "string", "minLength": 1, "maxLength": 300}, "our_runs": {"type": "array", "minItems": 2, "maxItems": 5, "items": {"type": "object", "additionalProperties": false, "required": ["run_index", "value", "unit"], "properties": {"run_index": {"type": "integer", "minimum": 1, "maximum": 5}, "value": {"type": ["number", "null"]}, "unit": {"type": "string", "minLength": 1, "maxLength": 40}, "reason_if_null": {"type": "string", "minLength": 0, "maxLength": 300}}}}, "comparable": {"type": "boolean"}, "difference_ratio": {"type": ["number", "null"]}, "verdict": {"type": "string", "enum": ["reproduced", "not_reproduced", "not_comparable", "unknown"]}, "reasons": {"type": "array", "minItems": 1, "maxItems": 8, "items": {"type": "string", "minLength": 1, "maxLength": 300}}}}, "rules": ["Two runs at least.", "not_comparable when the setup differs in a way that matters."], "status": "proposal", "admission": "not_admitted", "content_kind": "authored_guidance", "license": "MIT", "provenance": {"source": "original ta-amps authored proposal", "license": "MIT", "generator": "scripts/build_packs.py", "independently_verified": false}, "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": []}},
{"store": "seedpack.benchmark-scores", "key": "benchmark-scores.v0.1.0.template.staleness-report", "expected_version": 0, "producer_surface": "surface:tessera", "visibility_scope": "surface_only", "allowed_surfaces": [], "target_cohorts": ["all"], "tags": ["category:noun.eval_run", "pack:benchmark-scores", "role:template", "content:authored", "spdx:MIT", "status:quarantined"], "value": {"category_id": "noun.eval_run", "role": "template", "pack": "benchmark-scores", "pack_version": "0.1.0", "title": "Template: the staleness flags for every watched score as of one date", "category_reason": "A staleness report is a deterministic check run.", "describes": "the staleness flags for every watched score as of one date", "schema": {"type": "object", "additionalProperties": false, "required": ["as_of", "items", "proposed_refreshes"], "properties": {"as_of": {"type": "string", "pattern": "^[0-9]{4}-[0-9]{2}-[0-9]{2}", "maxLength": 40}, "items": {"type": "array", "minItems": 1, "maxItems": 500, "items": {"type": "object", "additionalProperties": false, "required": ["score_ref", "flags", "status"], "properties": {"score_ref": {"type": "string", "minLength": 1, "maxLength": 300}, "flags": {"type": "array", "minItems": 0, "maxItems": 5, "items": {"type": "string", "enum": ["older_than_max_age", "subject_changed", "benchmark_changed", "contradicted_by_our_run", "missing_setup"]}}, "status": {"type": "string", "enum": ["current", "stale", "contradicted", "unverified"]}}}}, "proposed_refreshes": {"type": "integer", "minimum": 0}}}, "rules": ["Flags come from the staleness policy only."], "status": "proposal", "admission": "not_admitted", "content_kind": "authored_guidance", "license": "MIT", "provenance": {"source": "original ta-amps authored proposal", "license": "MIT", "generator": "scripts/build_packs.py", "independently_verified": false}, "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": []}},
{"store": "seedpack.benchmark-scores", "key": "benchmark-scores.v0.1.0.schedule.daily-refresh", "expected_version": 0, "producer_surface": "surface:tessera", "visibility_scope": "surface_only", "allowed_surfaces": [], "target_cohorts": ["all"], "tags": ["category:noun.seed_hypothesis", "pack:benchmark-scores", "role:schedule_proposal", "content:authored", "spdx:MIT", "status:quarantined"], "value": {"category_id": "noun.seed_hypothesis", "role": "schedule_proposal", "pack": "benchmark-scores", "pack_version": "0.1.0", "title": "Proposed daily score refresh (not activated)", "category_reason": "A schedule proposal is stored as a seed hypothesis, following the research seed precedent.", "activation": "not_activated", "activation_path": "Recurring execution needs Tessera's flag-gated schedule support, which is being designed now. No scheduler is activated by this pack, and a schedule noun cannot authorize work by itself.", "cadence": {"every": "P1D", "at_utc": "06:00", "start": null, "stop": null, "start_reason": "Set by the owner when an authorized schedule adapter exists."}, "enqueues": ["benchmark-scores.refresh-scores", "benchmark-scores.flag-stale"], "per_window": {"max_sources": 20, "max_fetches": 40, "max_outstanding": 2, "max_runtime_seconds": 1200}, "enqueue_identity": "SHA-256 of the schedule key, its version and the window start, so duplicate ticks never create duplicate tasks.", "missed_windows": "Skip missed windows; never catch up more than one window.", "follow_up": "flag-stale proposes refresh or comparison verbs for flagged scores; an authorized requester enqueues those.", "status": "proposal", "admission": "not_admitted", "content_kind": "authored_guidance", "license": "MIT", "provenance": {"source": "original ta-amps authored proposal", "license": "MIT", "generator": "scripts/build_packs.py", "independently_verified": false}, "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.benchmark-scores/benchmark-scores.v0.1.0.staleness-policy", "target_category": "noun.reflex_procedure", "note": ""}]}}
]}