| name | health-check |
| description | Use this skill when the user wants to check the health of the team context, find stale docs, missing fields, or template compliance issues. Triggered by "health check", "check repo health", "stale docs", "team context health", "are our docs up to date", "what needs updating", "doc hygiene", "review system docs". Good to run during sprint retros or after major infrastructure changes. |
Team Context Health Check
Scans the team context for staleness, template drift, and missing content. Reports what needs attention so the team can keep the knowledge base accurate.
What to Check
1. System Doc Staleness
Read all files in systems/owned/ and systems/reference/. For each file:
- Look for
<!-- last-reviewed: YYYY-MM-DD --> on the first line
- If missing: flag as missing review date
- If present: calculate days since last review. Flag as stale if older than 90 days
Present results as a table:
System Doc Health:
| System | Last Reviewed | Status |
|--------|--------------|--------|
| ai-brain.md | 2026-03-31 | OK (0 days ago) |
| airflow.md | 2025-12-15 | STALE (107 days ago) |
| tableau.md | (missing) | NO DATE |
2. TODO / Placeholder Scan
Search all files in systems/ and references/ for:
<!-- TODO comments
- Empty sections (header followed immediately by another header or end of file)
- Placeholder text like
TBD, TODO, FIXME
Flag each finding with file path and line number.
3. Template Compliance
For each system doc, check which template it should follow:
- If the doc's Pointers section says "Investment: Maintain Only" -> light template
- Otherwise -> full template
Check required sections are present:
Full template required sections: Overview, Repos, Architecture, Key Entry Points, Related Systems, Pointers
Light template required sections: Overview, What the Team Owns, What the Team Does NOT Own, When It Breaks, Related Systems, Pointers
Flag missing sections.
4. Reference Data Freshness
Check references/monday_boards.md for the "last verified" date. Flag if older than 90 days.
5. Roster Integrity
Check references/team.md for completeness:
- Every team member has a non-empty Slack ID
- Every team member has a GitHub login
references/slack.md still points to team.md for the roster (link integrity)
Output Format
Group findings by severity:
## Team Context Health Report
### Needs Immediate Attention
- [list of critical issues: missing dates, TODOs in critical systems, template violations]
### Should Fix Soon
- [list of stale docs, minor template drift]
### Looking Good
- [count of healthy docs, compliant templates]
### Summary
X system docs checked, Y healthy, Z need attention.
Last full health check: [today's date]
After the Check
Suggest specific next steps:
- For stale docs: "Want me to open each one so you can review and update the
last-reviewed date?"
- For TODOs: "Want me to try filling these in, or should we create tasks for them?"
- For template violations: "Want me to restructure these docs to match the template?"
Don't auto-fix anything - present findings and let the user decide what to address.