xiduzo/wtf / skills/wtf.health
wtf.health
VerifiedThis skill should be used when a developer or lead wants a project status overview — what is blocked, stale, unverified, or waiting on a next step — for example "project health check", "what's blocking the release", "standup summary", or "what's left in this epic". Scans all open WTF issues — Traces plus legacy Tasks — and produces a triage-ready health report with actionable next steps.
Health
Project health dashboard.
This skill gives a full-stack view of open WTF issues in under a minute. It reports what is blocked, what is stale, and the clear next action for each problem found.
Process
0. GitHub CLI setup
Run steps 1–2 of ../references/gh-setup.md.
Stop if gh is not installed or not authenticated.
1. Choose the scope
Call AskUserQuestion (per ../references/questioning-style.md):
- question: "What scope do you want to check?"
- header: "Scope"
- options:
- Full project → all open Epics, Features, and Traces
- One Epic → all work under a specific Epic
- One Feature → all Traces under a specific Feature
For Epic or Feature scope, prompt for the issue number with options pre-filled from recent open issues.
2. Fetch all open issues
Run in parallel:
# Resolve $WTF_CLASS once — see ../references/issue-classification.md.
# labels mode → query by kind label; types mode → query by native issue type.
# The Trace query also matches legacy Task issues (Legacy Task reads in
# issue-classification.md) so migrated repos stay visible.
if [ "$WTF_CLASS" = types ]; then
gh issue list --search 'type:"Epic" state:open' --json number,title,labels,updatedAt --limit 50
gh issue list --search 'type:"Feature" state:open' --json number,title,body,labels,updatedAt --limit 100
gh issue list --search 'state:open (type:"Trace" OR type:"Task")' --json number,title,labels,updatedAt --limit 200
gh issue list --search 'type:"Bug" state:open' --json number,title,labels,updatedAt --limit 50
else
gh issue list --label "epic" --state open --json number,title,labels,updatedAt --limit 50
gh issue list --label "feature" --state open --json number,title,body,labels,updatedAt --limit 100
gh issue list --search 'state:open (label:trace OR label:task)' --json number,title,labels,updatedAt --limit 200 # --label "a,b" is AND
gh issue list --label "bug" --state open --json number,title,labels,updatedAt --limit 50
fi
The Feature query requests body because step 3 reads each Feature's User Stories and Trace Plan.
Also fetch open PRs to detect traces with an open PR but no verified label:
gh pr list --state open --json number,title,headRefName,body --limit 50
Resolve the delivery mode once per ../references/branch-setup.md ("Resolve the delivery mode"). A Feature's Delivery Override wins for that Feature. Step 3 reads it for the Feature-completion rows.
3. Classify issues into health categories
For each issue, check its labels against the expected lifecycle:
Epics:
| Signal | Category |
|---|---|
| No child Features linked | ⚠️ Epic has no Features |
| All child Features closed | ✅ Epic complete — needs wtf.retro |
Features:
| Signal | Category |
|---|---|
No designed label, no child Traces | ⚠️ Feature not designed, no traces |
| No Skeleton entry — Trace Plan absent, or item 1 is not the Skeleton | ⚠️ No Skeleton — the Spine is unproven |
| A story other than the primary has no Extension entry | ⚠️ Story not on the Spine — re-aim the Trace Plan |
| Partition gap — a story scenario in the Feature body that no Trace Plan entry claims | ⚠️ Unclaimed scenarios — re-aim the Trace Plan |
Has child Traces, none implemented | 🔵 In progress |
All child Traces verified, every Trace PR merged, Feature still open — staged delivery | ✅ Feature complete — needs Feature PR |
All child Traces verified, every Trace PR merged, Feature still open — trunk delivery | ⚠️ Feature not closed — the final Trace PR lacked Closes #<feature> |
Read each Feature body's User Stories and Trace Plan for the Skeleton, Extension, and partition checks. The primary story is the story named by the Skeleton entry. Compare the scenario names per story against the Scenario Claims in the Trace Plan entries. Do not check the synced scenario copies in Trace bodies — wtf.verify-trace owns drift detection.
Traces (legacy Task issues share this table — mark each such finding "legacy Task"):
| Signal | Category |
|---|---|
No designed label | ⚠️ Not designed — blocked before implement |
designed but not implemented for > 7 days | 🕐 Stale — may be forgotten |
implemented but not verified | ⏳ Waiting for QA |
verified but no open or merged PR | ⏳ Waiting for PR |
verified but its Trace Plan entry is unchecked | ⚠️ Trace Plan out of sync — tick the entry |
Has open PR but not verified | ⚠️ PR open without QA sign-off |
implemented + verified + PR merged | ✅ Done |
Bugs:
| Signal | Category |
|---|---|
| Open, no Trace linked in body | ⚠️ Bug not linked to a Trace |
| Open > 14 days | 🕐 Stale bug |
Staleness threshold for traces is 7 days since last update. For bugs, it is 14 days. These are heuristics — flag but do not auto-close anything.
4. Render the health report
Project Health — <scope> — <YYYY-MM-DD>
────────────────────────────────────────────────────────
Epics: [n open]
Features: [n open]
Traces: [n open] ([n] implemented, [n] verified, [n] stale, [n] legacy Tasks)
Bugs: [n open] ([n] stale)
────────────────────────────────────────────────────────
⚠️ Needs attention ([n])
[#n] Trace: <title>
Label gap: designed — not yet implemented (stale [n days])
Next: run implement-trace
[#n] Trace: <title>
implemented — not verified
Next: run verify-trace
[#n] Feature: <title>
Partition gap: scenario "<name>" claimed by no Trace Plan entry
Next: run refine to re-aim the Trace Plan
[#n] Feature: <title>
All traces verified — Feature PR not opened
Next: run create-pr from the feature branch targeting main (staged) — or, in trunk delivery, open a PR that carries `Closes #<n>`
[#n] Bug: <title>
Open [n days], no linked Trace
Next: run hotfix for a production fix, or refine the Feature so a Deepening Trace claims the failing scenario
🕐 Stale ([n])
[#n] Trace: <title>
Last updated [n days ago] — no activity since designed
────────────────────────────────────────────────────────
✅ Clean
[n] traces verified and merged
[n] features closed
If everything is clean, print: "All open issues are in a healthy state. Nothing needs attention."
5. Offer to act on findings
If findings exist, call AskUserQuestion (per ../references/questioning-style.md):
- question: "Would you like to act on any of these findings now?"
- header: "Next action"
- options:
- One option per ⚠️ finding (e.g. Verify Trace #42 → run
verify-trace) - Done — just the report → exit
- One option per ⚠️ finding (e.g. Verify Trace #42 → run
Route each selection to the appropriate skill with the issue number pre-loaded as context.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |