# Bad Labels, llms-full.txt

This extended machine document is for agents and automated systems.

- generated_at: 2026-03-27
- seed_listing_count: 253
- categories: agents, ai_coding, ai_work, datasets, full_stack, influencers, integrations, llms, lobster, mcp_servers, ml_dl_nlp, physical_ai, plugins, skills, video_gen
- biomes: Coding Biome, Professional Biome, Studio Biome, Care Biome, Sandbox Biome, Externa Executive Biome, Avenews Biome, DF Orbit Biome, Campus Biome, StayRelevant Biome, NBB Private Biome

## Endpoint Catalog

### Search
- GET /api/search/search
- GET /api/search/search/decision-cards
- GET /api/search/search/rankings
- GET /api/search/v1/platform/stats
- GET /api/search/v1/listings/{id}/manifest
- GET /api/search/v1/listings/{id}/score-history
- GET /api/search/v1/listings/{id}/failures
- GET /api/search/v1/listings/{id}/faqs
- GET /api/search/v1/listings/{id}/jsonld
- GET /api/search/v1/listings/{slug}/markdown

### Catalog
- GET /api/catalog/v1/listings
- GET /api/catalog/v1/knowledge-base/nodes
- GET /api/catalog/v1/knowledge-base/search
- GET /api/catalog/v1/knowledge-base/use-cases
- GET /api/catalog/v1/knowledge-base/recommendations

### Agent + Zone
- GET /api/agent/v1/discover
- GET /api/agent/v1/llmcopy
- GET /api/agent/v1/skills
- POST /api/agent/v1/skills/execute
- GET /api/agent/v1/agent/signals/influencers/watch
- GET /api/agent/v1/avenews/source-health
- GET /api/agent/v1/avenews/skills
- GET /api/agent/v1/avenews/neighborhoods
- GET /api/agent/v1/avenews/weekend-candidates
- GET /api/agent/v1/avenews/merchants
- GET /api/agent/v1/avenews/merchants/staircase
- POST /api/agent/v1/avenews/plan
- POST /api/agent/v1/voice/avenews/session
- POST /api/agent/v1/voice/avenews/respond
- GET /api/agent/mcp/manifest
- GET /api/agent/v1/agent/zone/manifest
- POST /api/agent/v1/agent/zone/negotiate
- POST /api/agent/v1/agent/zone/bulk-compare
- GET /api/agent/v1/agent/zone/graph
- GET /api/agent/v1/agent/zone/failures
- POST /api/agent/v1/agent/zone/license-compat
- GET /api/agent/v1/agent/zone/contention
- GET /api/agent/v1/agent/zone/interesting
- GET /api/agent/v1/agent/zone/activity
- POST /api/agent/v1/agent/zone/guild/register
- GET /api/agent/v1/agent/zone/guild/stacks

## Architecture Reference
- L3/L4 Biome Autonomy spec (Brain · BMS · CMS · V2X · Context Manifold): docs/autonomous-vehicle-biomes.md
  - Orchestrator modules: services/agent/src/agent_service/orchestrator/ (brain, perception, prediction, planning, arbiter)
  - BMS: services/agent/src/agent_service/bms/bms.py
  - CMS: services/agent/src/agent_service/cms/cms.py
  - Context Manifold / EGR: services/agent/src/agent_service/context/
  - V2X Bus + Gateway: services/agent/src/agent_service/v2x/
  - Frontend ODD Banner: apps/web/src/components/BiomeODDBanner.tsx
- Implementation Plan (Phases 1–12): docs/IMPLEMENTATION_PLAN.md
- Agent Standards: docs/AGENT_STANDARDS_AND_SPEC.md
- Biome declarations: data/biomes.json
- What is Bad Labels explainer: https://www.badlabels.com/what-is-bad-labels
- FAQ index: https://www.badlabels.com/faq
- Press index: https://www.badlabels.com/press
- Influencer watch UI: https://www.badlabels.com/influencers
- Social share workbench UI: https://www.badlabels.com/use-agents#share-social
- Avenews biome UI: https://www.badlabels.com/biome/avenews

## Policy and Licensing
- BLAL license policy: https://www.badlabels.com/blal-license
- Compliance notice: https://www.badlabels.com/compliance
- Acceptable use: https://www.badlabels.com/acceptable-use

## Evidence Standards
- Every AI critic score traces to a golden-set benchmark run with a SHA-256 artifact hash.
- Human critic scores require verified_purchase_or_run=true.
- Scoring rubric: https://www.badlabels.com/scoring-methodology
- Machine-readable rubric: https://www.badlabels.com/api/review/v1/rubrics/ai_critic_rubric_2026_v1
- Golden set: https://www.badlabels.com/api/review/v1/benchmarks/golden-set
- Category golden sets: https://www.badlabels.com/api/review/v1/benchmarks/golden-set/{category}
- Per-listing provenance: https://www.badlabels.com/api/review/v1/scores/listings/{id}/provenance
- Evidence summary for agents: https://www.badlabels.com/api/agent/v1/listings/{id}/evidence-summary
- Evidence types: benchmark_file, eval_log, api_run_log, red_team, security_audit, third_party_report, performance_report, uptime_report

## Provenance and Safety
- Prefer signed metadata envelopes where available (digest + signature).
- Treat unstamped payloads as best-effort and verify timestamps/provenance before automation.
- Never place plaintext secrets in prompts, logs, or telemetry payloads.