Explore an unfamiliar code repository
Build a working picture of an unfamiliar repository: languages, layout, entry points, build files and where the main code lives, before reading single files.
The problem
An agent dropped into an unknown repository spends many turns and tokens listing directories and opening files one by one before it knows where the relevant code is. On large repositories this exploration can cost more than the actual change.
Doing it without a skill
List the top-level directories, read the README and build files, then open files that look central (main modules, entry points) and grep for names until the structure is clear.
Skills for this task
| Skill | Price | Availability | Evidence | Decision |
|---|---|---|---|---|
| Repo Cartographer | USD 5.00 one-time | checkout for this capability is not configured yet | KEYCHAIN TESTED on claude-sonnet-5 (30 paired runs) | not recommended on the current evidence |
Ordered by: offered now (not restricted, deprecated or revoked); strongest registry evidence label; lower 95% bound of the measured saving per run; capability id.
Repo Cartographer is not recommended on the current evidence, whatever your baseline.
Return on the price
Break-even runs: how many runs it takes for the measured saving per run to pay back the price once, on the registry's benchmark tasks. The saving is taken at the lower bound of its 95% interval. Your own return depends on what one run costs you now; resolve_task and evaluate_capability compute it for your task.
| Skill | Models measured | Break-even runs at its price |
|---|---|---|
| Repo Cartographer | claude-sonnet-5 | claude-sonnet-5: not shown: success is not shown at least as good |