| name | create-feature-catalog |
| description | Deep-mine the source code to discover implemented features; writes ai/analysis/FEATURE_CATALOG.md. |
| disable-model-invocation | true |
Build the feature catalog — the highest-value artifact for agents. Budget significant
exploration; use repo-explorer for the heavy reading.
Method
- Start from user-visible surfaces: routes, UI entry points, CLI commands, public
APIs. Each surface is a candidate feature.
- For each feature, trace the touch list across layers: UI, backend/services,
persistence (tables/collections/files), and tests that exercise it.
- Cluster and name features the way a USER would name them, not by module names.
Output — ai/analysis/FEATURE_CATALOG.md
For every feature: name · business goal (one line) · touch list per layer ·
verifying tests · related features. End with two sections agents use most:
- "Where new code lives" — a decision tree from feature-type to target directories.
- The 3-file rule — for each feature, the 3 files to read first to understand it.
Rules
- Every entry
[inferred]. Where a layer can't be confirmed, write "UNSURE".
- Do not modify source. Update
ai/guide/FEATURE_MAP.md candidate list to reference
the catalog, nothing more.
- Print a sampling guide at the end: the 5 entries a human should spot-check first
(pick the ones you are least sure of).
