Wtf.health

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 Tasks
  • One Epic → all work under a specific Epic
  • One Feature → all Tasks 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 $WTF_CLASS once — see ../references/issue-classification.md.
# labels mode → query by kind label; types mode → query by native issue type.
if [ "$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,labels,updatedAt --limit 100
  gh issue list --search 'type:"Task" state:open'    --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,labels,updatedAt --limit 100
  gh issue list --label "task"    --state open --json number,title,labels,updatedAt --limit 200
  gh issue list --label "bug"     --state open --json number,title,labels,updatedAt --limit 50
fi

Also fetch open PRs to detect tasks with an open PR but no verified label:

gh pr list --state open --json number,title,headRefName,body --limit 50

3. Classify issues into health categories

For each issue, check its labels against the expected lifecycle:

Epics:

Features:

Tasks:

Bugs:

Staleness threshold for tasks 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]
Tasks:    [n open]  ([n] implemented, [n] verified, [n] stale)
Bugs:     [n open]  ([n] stale)

────────────────────────────────────────────────────────

⚠️  Needs attention ([n])

  [#n] Task: <title>
       Label gap: designed — not yet implemented (stale [n days])
       Next: run implement-task

  [#n] Task: <title>
       implemented — not verified
       Next: run verify-task

  [#n] Feature: <title>
       All tasks verified — Feature PR not opened
       Next: run create-pr targeting main

  [#n] Bug: <title>
       Open [n days], no linked task
       Next: run write-task to create a fix task

🕐  Stale ([n])

  [#n] Task: <title>
       Last updated [n days ago] — no activity since designed

────────────────────────────────────────────────────────

✅  Clean
  [n] tasks 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 Task #42 → run verify-task)
  • Done — just the report → exit

Route each selection to the appropriate skill with the issue number pre-loaded as context.