{"format": "ta-amps.seed-pack.nouns.v1", "pack_id": "multimodal-mask-and-transcript-alignment", "status": "proposal_only", "records": [
{"store": "seedpack.multimodal-mask-and-transcript-alignment", "key": "multimodal-mask-and-transcript-alignment.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:multimodal-mask-and-transcript-alignment", "role:pack", "content:authored", "spdx:MIT", "status:quarantined"], "value": {"category_id": "noun.skill_pack", "role": "pack", "pack": "multimodal-mask-and-transcript-alignment", "pack_version": "0.1.0", "title": "A collection of nouns and verbs for aligning segmentation masks, speech word timings, and layout safe zones", "lets_workers": "Normalize pixel or polygon segmentation masks, group ASR word timings into caption utterances, check audio-visual cue drift in milliseconds, and select non-occluding layout safe zones.", "not_for": "Fabricating speech transcripts or overriding occlusion conflicts silently.", "flow": [{"verb": "multimodal-mask-and-transcript-alignment.normalize-masks-and-utterances", "when": "First: normalize mask boxes and group word timings into utterances."}, {"verb": "multimodal-mask-and-transcript-alignment.audit-audiovisual-cue-drift", "when": "Check millisecond sync between spoken object cues and mask onset."}, {"verb": "multimodal-mask-and-transcript-alignment.route-non-occluding-safe-zones", "when": "Assign each utterance to the highest-priority non-occluding safe zone."}], "guardrails": [{"store": "seedpack.multimodal-mask-and-transcript-alignment", "key": "multimodal-mask-and-transcript-alignment.v0.1.0.guardrail.non-occluding-captions", "version": 1, "value_sha256": "5276f1d00933862756385ba8d436a9565ea786d7d243ff3f5e7fcecec4af2283"}, {"store": "seedpack.multimodal-mask-and-transcript-alignment", "key": "multimodal-mask-and-transcript-alignment.v0.1.0.guardrail.av-drift-budget", "version": 1, "value_sha256": "ba6b7a57375d42fad341647becb92864ca957c8da692ea662188e90932505a5e"}], "tool": {"store": "seedpack.multimodal-mask-and-transcript-alignment", "key": "multimodal-mask-and-transcript-alignment.v0.1.0.tool.multimodal-align", "version": 1, "value_sha256": "c68f8f68175232b28f6a8f6342db919fe93560095a02725bb7a492a5c8128e1d"}, "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.multimodal-mask-and-transcript-alignment/multimodal-mask-and-transcript-alignment.v0.1.0.method", "target_category": "noun.skill_pack", "note": ""}, {"relation": "requires_noun", "target_id": "seedpack.multimodal-mask-and-transcript-alignment/multimodal-mask-and-transcript-alignment.v0.1.0.guardrail.non-occluding-captions", "target_category": "noun.skill_pack", "note": ""}, {"relation": "requires_noun", "target_id": "seedpack.multimodal-mask-and-transcript-alignment/multimodal-mask-and-transcript-alignment.v0.1.0.guardrail.av-drift-budget", "target_category": "noun.skill_pack", "note": ""}, {"relation": "requires_noun", "target_id": "seedpack.multimodal-mask-and-transcript-alignment/multimodal-mask-and-transcript-alignment.v0.1.0.rules.mask-intrusion-and-sync-math", "target_category": "noun.reflex_procedure", "note": ""}, {"relation": "requires_noun", "target_id": "seedpack.multimodal-mask-and-transcript-alignment/multimodal-mask-and-transcript-alignment.v0.1.0.tool.multimodal-align", "target_category": "noun.code_node", "note": ""}, {"relation": "requires_noun", "target_id": "seedpack.multimodal-mask-and-transcript-alignment/multimodal-mask-and-transcript-alignment.v0.1.0.template.alignment-report", "target_category": "noun.eval_run", "note": ""}]}},
{"store": "seedpack.multimodal-mask-and-transcript-alignment", "key": "multimodal-mask-and-transcript-alignment.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:multimodal-mask-and-transcript-alignment", "role:method", "content:authored", "spdx:MIT", "status:quarantined"], "value": {"category_id": "noun.skill_pack", "role": "method", "pack": "multimodal-mask-and-transcript-alignment", "pack_version": "0.1.0", "title": "Method: normalize coordinates, verify A/V cue drift, then route safe zones", "category_reason": "Shared geometric and temporal procedure for multimodal layout alignment.", "lifecycle": ["normalize-masks-and-utterances converts pixel or polygon masks to [0..1] boxes and groups words by pause/length.", "audit-audiovisual-cue-drift computes (word_start_s - mask_onset_s) * 1000 for matched labels.", "route-non-occluding-safe-zones picks the highest-priority zone with intrusion <= max_intrusion and runs alignment-checks."], "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.multimodal-mask-and-transcript-alignment", "key": "multimodal-mask-and-transcript-alignment.v0.1.0.guardrail.non-occluding-captions", "expected_version": 0, "producer_surface": "surface:tessera", "visibility_scope": "surface_only", "allowed_surfaces": [], "target_cohorts": ["all"], "tags": ["category:noun.skill_pack", "pack:multimodal-mask-and-transcript-alignment", "role:guardrail", "content:authored", "spdx:MIT", "status:quarantined"], "value": {"category_id": "noun.skill_pack", "role": "guardrail", "pack": "multimodal-mask-and-transcript-alignment", "pack_version": "0.1.0", "title": "Guardrail: reroute captions away from segmented subjects and HUD exclusions", "rule": "Any candidate safe zone whose intersection area divided by zone area exceeds max_intrusion (default 0.05) is disqualified if a cleaner zone exists, or flagged with occlusion_conflict=true.", "how_it_is_checked": "align_multimodal_timeline evaluates mask_intrusion_ratio across all active masks and HUD boxes.", "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.multimodal-mask-and-transcript-alignment", "key": "multimodal-mask-and-transcript-alignment.v0.1.0.guardrail.av-drift-budget", "expected_version": 0, "producer_surface": "surface:tessera", "visibility_scope": "surface_only", "allowed_surfaces": [], "target_cohorts": ["all"], "tags": ["category:noun.skill_pack", "pack:multimodal-mask-and-transcript-alignment", "role:guardrail", "content:authored", "spdx:MIT", "status:quarantined"], "value": {"category_id": "noun.skill_pack", "role": "guardrail", "pack": "multimodal-mask-and-transcript-alignment", "pack_version": "0.1.0", "title": "Guardrail: block render readiness when spoken cues drift beyond 250 ms from mask onset", "rule": "When abs(av_offset_ms) > max_drift_ms, av_drift_violations increments and ready_for_render becomes false.", "how_it_is_checked": "align_multimodal_timeline records every av_sync_events entry with drift_exceeded.", "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.multimodal-mask-and-transcript-alignment", "key": "multimodal-mask-and-transcript-alignment.v0.1.0.rules.mask-intrusion-and-sync-math", "expected_version": 0, "producer_surface": "surface:tessera", "visibility_scope": "surface_only", "allowed_surfaces": [], "target_cohorts": ["all"], "tags": ["category:noun.reflex_procedure", "pack:multimodal-mask-and-transcript-alignment", "role:rules", "content:authored", "spdx:MIT", "status:quarantined"], "value": {"category_id": "noun.reflex_procedure", "role": "rules", "pack": "multimodal-mask-and-transcript-alignment", "pack_version": "0.1.0", "title": "Deterministic formulas for box IoU, polygon Shoelace area, intrusion ratio, and A/V offset", "intrusion_ratio": "intersection_area(zone_box, mask_box) / area(zone_box).", "polygon_conversion": "Tight [min_x, min_y, max_x - min_x, max_y - min_y] box plus half-sum Shoelace area.", "av_offset_ms": "round((word_start_s - first_mask_onset_s) * 1000.0, 2).", "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.multimodal-mask-and-transcript-alignment", "key": "multimodal-mask-and-transcript-alignment.v0.1.0.template.normalized-timeline", "expected_version": 0, "producer_surface": "surface:tessera", "visibility_scope": "surface_only", "allowed_surfaces": [], "target_cohorts": ["all"], "tags": ["category:noun.scene_recipe", "pack:multimodal-mask-and-transcript-alignment", "role:template", "content:authored", "spdx:MIT", "status:quarantined"], "value": {"category_id": "noun.scene_recipe", "role": "template", "pack": "multimodal-mask-and-transcript-alignment", "pack_version": "0.1.0", "title": "Template: normalized multimodal mask and utterance timeline", "describes": "utterance spans and [0..1] normalized mask frames for one media clip", "schema": {"type": "object", "additionalProperties": false, "required": ["clip_id", "utterance_count", "mask_frame_count"], "properties": {"clip_id": {"type": "string", "minLength": 1, "maxLength": 120}, "utterance_count": {"type": "integer", "minimum": 0}, "mask_frame_count": {"type": "integer", "minimum": 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.multimodal-mask-and-transcript-alignment", "key": "multimodal-mask-and-transcript-alignment.v0.1.0.template.alignment-report", "expected_version": 0, "producer_surface": "surface:tessera", "visibility_scope": "surface_only", "allowed_surfaces": [], "target_cohorts": ["all"], "tags": ["category:noun.eval_run", "pack:multimodal-mask-and-transcript-alignment", "role:template", "content:authored", "spdx:MIT", "status:quarantined"], "value": {"category_id": "noun.eval_run", "role": "template", "pack": "multimodal-mask-and-transcript-alignment", "pack_version": "0.1.0", "title": "Template: checked multimodal safe-zone and sync alignment report", "describes": "safe-zone placements, intrusion ratios, A/V drift counts, and render-readiness flag", "schema": {"type": "object", "additionalProperties": false, "required": ["clip_id", "alignment_plan_sha256", "occlusion_conflicts", "av_drift_violations", "ready_for_render", "placements"], "properties": {"clip_id": {"type": "string", "minLength": 1, "maxLength": 120}, "alignment_plan_sha256": {"type": "string", "pattern": "^[0-9a-f]{64}$"}, "occlusion_conflicts": {"type": "integer", "minimum": 0}, "av_drift_violations": {"type": "integer", "minimum": 0}, "ready_for_render": {"type": "boolean"}, "placements": {"type": "array", "minItems": 1, "maxItems": 500, "items": {"type": "object", "additionalProperties": false, "required": ["utterance_id", "selected_zone_id", "selected_bbox", "max_intrusion_ratio", "occlusion_conflict"], "properties": {"utterance_id": {"type": "string", "minLength": 1, "maxLength": 40}, "selected_zone_id": {"type": "string", "minLength": 1, "maxLength": 40}, "selected_bbox": {"type": "array", "minItems": 4, "maxItems": 4, "items": {"type": "number", "minimum": 0, "maximum": 1}}, "max_intrusion_ratio": {"type": "number", "minimum": 0, "maximum": 1}, "occlusion_conflict": {"type": "boolean"}}}}}}, "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.multimodal-mask-and-transcript-alignment", "key": "multimodal-mask-and-transcript-alignment.v0.1.0.alignment-checks", "expected_version": 0, "producer_surface": "surface:tessera", "visibility_scope": "surface_only", "allowed_surfaces": [], "target_cohorts": ["all"], "tags": ["category:noun.benchmark_suite", "pack:multimodal-mask-and-transcript-alignment", "role:record_checks", "content:authored", "spdx:MIT", "status:quarantined"], "value": {"category_id": "noun.benchmark_suite", "role": "record_checks", "pack": "multimodal-mask-and-transcript-alignment", "pack_version": "0.1.0", "title": "Deterministic checks for multimodal alignment reports", "fixture_kind": "finding_checks", "purpose": "Deterministic checks every multimodal mask and transcript alignment report must pass.", "deterministic_checks": [{"id": "plan-digest-pinned", "rule": {"rule": "pattern_present", "path": "alignment_plan_sha256", "pattern": "^[0-9a-f]{64}$"}, "why": "Every alignment report pins the SHA-256 digest of its placement schedule."}, {"id": "placements-complete", "rule": {"rule": "each_has", "path": "placements", "fields": ["utterance_id", "selected_zone_id", "selected_bbox"], "min": 1}, "why": "Every utterance placement names its utterance_id, selected_zone_id, and bounding box."}, {"id": "hosts-local", "rule": {"rule": "hosts_local_only", "path": "$"}, "why": "Alignment reports operate on local media timelines without external hosts."}], "golden_fixtures": [{"id": "good-alignment-report", "label": "known_good", "finding": {"clip_id": "synth-clip-01", "alignment_plan_sha256": "f0e1d2c3b4a5968778695a4b3c2d1e0ff0e1d2c3b4a5968778695a4b3c2d1e0f", "ready_for_render": true, "placements": [{"utterance_id": "utt.001", "selected_zone_id": "top_band", "selected_bbox": [0.08, 0.08, 0.84, 0.14]}]}}, {"id": "bad-missing-zone", "label": "known_bad", "finding": {"clip_id": "synth-clip-02", "alignment_plan_sha256": "f0e1d2c3b4a5968778695a4b3c2d1e0ff0e1d2c3b4a5968778695a4b3c2d1e0f", "ready_for_render": false, "placements": [{"utterance_id": "utt.001", "selected_zone_id": "", "selected_bbox": []}]}}, {"id": "bad-missing-sha", "label": "known_bad", "finding": {"clip_id": "synth-clip-03", "alignment_plan_sha256": "not-a-sha", "ready_for_render": true, "placements": [{"utterance_id": "utt.001", "selected_zone_id": "bottom_band", "selected_bbox": [0.08, 0.78, 0.84, 0.14]}]}}], "expected": {"good-alignment-report": {"verdict": "pass", "failed_checks": []}, "bad-missing-zone": {"verdict": "fail", "failed_checks": ["placements-complete"]}, "bad-missing-sha": {"verdict": "fail", "failed_checks": ["plan-digest-pinned"]}}, "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.multimodal-mask-and-transcript-alignment", "key": "multimodal-mask-and-transcript-alignment.v0.1.0.tool.multimodal-align", "expected_version": 0, "producer_surface": "surface:tessera", "visibility_scope": "surface_only", "allowed_surfaces": [], "target_cohorts": ["all"], "tags": ["category:noun.code_node", "pack:multimodal-mask-and-transcript-alignment", "role:tool_card", "content:authored", "spdx:MIT", "status:quarantined"], "value": {"category_id": "noun.code_node", "role": "tool_card", "pack": "multimodal-mask-and-transcript-alignment", "pack_version": "0.1.0", "title": "Tool card: multimodal_align standard-library CLI and library", "name": "multimodal_align", "code": {"path": "packs/multimodal-mask-and-transcript-alignment/code", "files": 6, "sha256": "046a421118da85e62c51dfddc78f398c02121e94b99c34f3c384d83382756830"}, "runtime": "Python 3.10 or later, standard library only.", "commands": ["python3 -m multimodal_align align TIMELINE.json [--max-drift-ms 250] [--max-intrusion 0.05]"], "does": ["Normalizes bounding-box and polygon segmentation masks and groups word timestamps into utterances.", "Computes mask intrusion ratios, routes non-occluding caption safe zones, and flags A/V sync drift."], "does_not": ["Transcode media files, invoke neural weights, or open network connections."], "code_license": "MIT", "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": "verifies", "target_id": "seedpack.multimodal-mask-and-transcript-alignment/multimodal-mask-and-transcript-alignment.v0.1.0.rules.mask-intrusion-and-sync-math", "target_category": "noun.reflex_procedure", "note": ""}]}},
{"store": "seedpack.multimodal-mask-and-transcript-alignment", "key": "multimodal-mask-and-transcript-alignment.v0.1.0.source.w3c-webvtt", "expected_version": 0, "producer_surface": "surface:tessera", "visibility_scope": "surface_only", "allowed_surfaces": [], "target_cohorts": ["all"], "tags": ["category:noun.public_source", "pack:multimodal-mask-and-transcript-alignment", "role:source", "content:authored", "spdx:MIT", "status:quarantined"], "value": {"category_id": "noun.public_source", "role": "source", "pack": "multimodal-mask-and-transcript-alignment", "pack_version": "0.1.0", "title": "W3C WebVTT: The Web Video Text Tracks Format", "url": "https://www.w3.org/TR/webvtt1/", "checked_on": "2026-10-31", "provides": "Cue timing, region box, and viewport safe-area positioning semantics for timed captions.", "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.multimodal-mask-and-transcript-alignment", "key": "multimodal-mask-and-transcript-alignment.v0.1.0.source.itu-bt1359", "expected_version": 0, "producer_surface": "surface:tessera", "visibility_scope": "surface_only", "allowed_surfaces": [], "target_cohorts": ["all"], "tags": ["category:noun.public_source", "pack:multimodal-mask-and-transcript-alignment", "role:source", "content:authored", "spdx:MIT", "status:quarantined"], "value": {"category_id": "noun.public_source", "role": "source", "pack": "multimodal-mask-and-transcript-alignment", "pack_version": "0.1.0", "title": "ITU-R BT.1359-1: Relative timing of sound and vision for broadcasting", "url": "https://www.itu.int/rec/R-REC-BT.1359/", "checked_on": "2026-10-31", "provides": "Perceptual audio-visual lead and lag detectability thresholds informing drift gates.", "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.multimodal-mask-and-transcript-alignment", "key": "multimodal-mask-and-transcript-alignment.v0.1.0.example.safe-zone-reroute", "expected_version": 0, "producer_surface": "surface:tessera", "visibility_scope": "surface_only", "allowed_surfaces": [], "target_cohorts": ["all"], "tags": ["category:noun.eval_run", "pack:multimodal-mask-and-transcript-alignment", "role:example", "synthetic:true", "spdx:MIT", "status:quarantined"], "value": {"category_id": "noun.eval_run", "role": "example", "pack": "multimodal-mask-and-transcript-alignment", "pack_version": "0.1.0", "title": "Worked example: synthetic caption reroute from occluded bottom_band to top_band", "conforms_to": {"store": "seedpack.multimodal-mask-and-transcript-alignment", "key": "multimodal-mask-and-transcript-alignment.v0.1.0.template.alignment-report", "version": 1, "value_sha256": "dee16cb2d6c5195bb2ba810dc084f97348d174f60a3238cc62a928b7c883a6bc"}, "instance": {"clip_id": "synth-valve-demo", "alignment_plan_sha256": "f0e1d2c3b4a5968778695a4b3c2d1e0ff0e1d2c3b4a5968778695a4b3c2d1e0f", "occlusion_conflicts": 0, "av_drift_violations": 0, "ready_for_render": true, "placements": [{"utterance_id": "utt.001", "selected_zone_id": "top_band", "selected_bbox": [0.08, 0.08, 0.84, 0.14], "max_intrusion_ratio": 0.0, "occlusion_conflict": false}]}, "shows": "A foreground mask occupying [0.10, 0.70, 0.80, 0.25] triggers a clean reroute to top_band with zero intrusion.", "source": "Synthetic. Written for this pack.", "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": []}}
]}