# Production Source Authority Registry

Status: `pass`
Created: `2026-06-13T15:03:25Z`

## The Single Idea

AINA already has title, responsibility, workflow, evidence-atlas, PKM, and salvage
work spread across repos. This registry is the build-time map that prevents the
embedding lane from rediscovering weaker sources and accidentally embedding noisy
marketplace labels as semantic truth.

## Title Authority Metrics

- Jobs-research title audit rows: `37478`
- Trusted audit titles: `15104`
- Pass-draft clean-evidence titles: `6700`
- Pass-draft clean-evidence rows: `44440`
- Source-authority inventory assets: `46`
- Company/employer assets inventoried: `5`

## Sources

| Source | Consume as | Status | Path |
| --- | --- | --- | --- |
| `engine_room_harvest_source_map` | `build-time source inventory` | `present` | `/srv/aina/aina-data-engine-room/artifacts/validation/harvest_source_map_v1.json` |
| `engine_room_title_ledger` | `title precedence and lineage map` | `present` | `/srv/aina/aina-data-engine-room/docs/TITLE-LEDGER.md` |
| `engine_room_mapping_chain_ledger` | `title-to-role-to-workflow join map` | `present` | `/srv/aina/aina-data-engine-room/docs/MAPPING-CHAIN-LEDGER.md` |
| `cross_repo_salvage_map` | `prior-work navigation map` | `present` | `/home/ali/conductor/aina-consolidated/20-references/linear/doc__cross-repo-salvage-map.md` |
| `jobs_research_title_audit` | `title authority and SOC correction` | `present` | `/home/ali/conductor/repos/aina-jobs-research/project-summary-package/organized-outputs/02-audience-title-and-onet-workflow-pilot/outputs/outputs/audience_titles_enriched_audit.csv` |
| `jobs_research_clean_candidates_pass_draft` | `clean evidence retrieval candidate substrate` | `candidate_pass_draft_present` | `/home/ali/conductor/repos/aina-jobs-research/project-summary-package/organized-outputs/03-combined-source-intelligence-candidate-layer/outputs/outputs/combined_clean_evidence_candidates_v1.pass_draft.jsonl` |
| `jobs_research_source_intelligence_v1` | `role responsibility workflow tool affordance source package` | `present` | `/home/ali/conductor/repos/aina-jobs-research/project-summary-package/exports/source_intelligence_v1/manifest.json` |
| `aina_core_evidence_atlas_title_aliases` | `stitched title alias evidence` | `present` | `/home/ali/conductor/repos/aina-core/evidence/canonical/local_title_aliases.parquet` |
| `aina_core_evidence_atlas_responsibilities` | `stitched role responsibility evidence` | `present` | `/home/ali/conductor/repos/aina-core/evidence/canonical/role_responsibility_evidence.parquet` |
| `engine_room_source_truth_ledger` | `public baseline source terms and versions` | `present` | `/srv/aina/aina-data-engine-room/artifacts/sources/source_truth_ledger.json` |

## Anti-Regression Samples

| Input title | Embedding title | Status | Authority SOC |
| --- | --- | --- | --- |
| `j.p. morgan wealth management - private client advisor - tulsa ok` | `private client advisor` | `derived_clean_title` | `13-2052.00` |
| `seasonal sales associate - baybrook mall` | `seasonal sales associate` | `derived_clean_title` | `41-2031.00` |
| `lcur account manager - bellevue wa` | `account manager` | `derived_clean_title` | `11-2022.00` |
| `senior banker - short pump financial center` | `senior banker` | `derived_clean_title` | `41-3031.00` |
| `support associate` | `support associate` | `missing` | `43-4051` |

## Build Contract

`serviceable_title` and `semantic_review` chunks must preserve the raw market
title in metadata, but use the clean title and SOC authority from jobs-research
when that authority exists. Raw market rows and candidate/repair/hold rows remain
blocked from vector authority until converted into derived clean chunks.

---

Ali Mehdi Mukadam - co-authored with Codex