overviewlisted
Install: claude install-skill qupunto/wss
# The project at a glance
**This skill writes nothing.** No record, no checkpoint, no commit, no sweep.
Anything that looks wrong is reported with the flag that owns fixing it, in
one line, and left alone.
## Run the probe first
Every mechanical number in the report comes from one script, run from the
project directory:
```bash
S="${CLAUDE_CONFIG_DIR:-$HOME/.claude}"
[ -x "$S/wss/tests/wss-doctor.sh" ] || S=$(ls -d "$S"/plugins/cache/*/wss/*/ 2>/dev/null | tail -1)
bash "$S"/skills/overview/assets/wss-probe.sh
```
<!-- load-bearing: the only route to contracts that survives publication — see the decision log before removing -->
The two resolution lines are [`contracts`](../contracts/SKILL.md)'
canonical form.
The probe is read-only and offline. It resolves `.claude/WSS.WORKFLOW.json`
itself — conventional fallbacks, the `.claude/WSS.LANE` selector, `WSS.lanes.named`
overrides, all per [`WSS.MANIFEST.md`](../../wss/workflow/WSS.MANIFEST.md) — counts every
record, runs `wss-doctor.sh`, computes each sweep baseline's distance from HEAD
— by calling `wss-sweep-distance.sh` beside it, the one implementation of that
measurement, which `--wss-wrap` step 7 runs `--compact` for its closing line —
locates the roadmap's first goal with open blocks, and locates the release
list's first milestone not marked completed. **The roadmap is lane-resolved
and the release list never is** — one release checkpoint per project, however
many lanes it runs, per
[`WSS.RECORD-CONTRACT.md`](../../w