| name | self-verify |
| description | Use when the user says "check the UI", "verify changes", "does it look right", or after completing UI work — quick browser spot-check (not a deploy gate; that is /e2e-verify). |
| user-invocable | true |
| tier | flexible |
| kind | verification |
Run: `bash "$(git rev-parse --show-toplevel)/bin/harness-update-check"`
- `UPGRADE_AVAILABLE ` → tell the user: "agent-harness is available (you have ). Run `/harness-update` to pull it in." Then continue.
- `JUST_UPGRADED ` → tell the user: "agent-harness upgraded → ." Then continue.
- No output → continue silently.
Self-Verify (Browser)
Visually verify UI changes using the browser. Use when the user says "check the UI", "verify changes", "does it look right", or after completing UI work.
Steps
- Ensure dev server is running — reference the
dev-server skill if needed. Check port with lsof -i :${CONDUCTOR_PORT:-$HARNESS_DEV_PORT} (inside a Conductor workspace the dev server binds to CONDUCTOR_PORT; outside it uses HARNESS_DEV_PORT)
- Navigate to the relevant page (default:
http://localhost:${CONDUCTOR_PORT:-$HARNESS_DEV_PORT})
- Check browser console for errors or warnings
- Visually verify that changes render correctly
- Test basic interactions — click buttons, fill forms, navigate links
- Report findings — what looks correct, any issues found
Tools
Use Claude-in-Chrome MCP tools (mcp__claude-in-chrome__*) or Playwright (mcp__playwright__*) for browser automation.
Load tools with ToolSearch before first use.
What to Check
- Page renders without blank screens or layout breaks
- New UI elements appear in the correct location
- Interactive elements respond correctly (buttons, forms, dropdowns)
- No JavaScript errors in the browser console
- Navigation links work
- Data loads and displays correctly
Report Format
## Self-Verify Report
### Page: [URL]
**Rendering:** PASS/FAIL
**Interactions:** PASS/FAIL
**Console errors:** None / [list errors]
### Findings
- [finding]
### Verdict: PASS / NEEDS FIXES
