confab-statuslisted
Install: claude install-skill Anselmoo/werkstoff
Report the current state of confab's analysis for this repository: which
domain audits have run, whether their results are stale relative to the
latest commit, and the single most useful next action.
## Steps
1. Determine `repo_root`.
2. Run:
```
python3 "${CLAUDE_PLUGIN_ROOT}/scripts/status_dashboard.py" <repo_root>
```
This is read-only — it opens existing `analysis/confab/*_summary.json`
sidecars and `analysis/confab/ledger.json` if present, and runs `git log -1`
(read-only) to establish a staleness reference. It never runs an audit
itself. It writes `analysis/confab/findings_dashboard_data.json` and
`analysis/confab/reports/findings-dashboard.html`, and prints the same data
plus a `suggestion` field to stdout.
3. If `analysis/confab/ledger.json` exists (a `confab-cycle` has run at
least once), also render its pass history as a burndown chart —
separate from the snapshot dashboard above, purely additive:
```
python3 "${CLAUDE_PLUGIN_ROOT}/scripts/build_burndown_html.py" <repo_root> \
--template "${CLAUDE_PLUGIN_ROOT}/assets/burndown-viewer.html"
```
This reads only `ledger.json`'s already-persisted `passes` array and
current `findings` statuses — it does not re-derive or estimate
anything not already on disk. Writes
`analysis/confab/reports/BURNDOWN.html`. Skip this step entirely if no
ledger exists yet.
4. Report to the user, per domain: has it ever run, how many findings,
and whether it's stale relative