agentleFS
stoa-platformstoa-platform/stoa/.claude/skills/verify-mega/SKILL.md

verify-mega

Skill9 starsChanged 7 months agoApache-2.0View on GitHub

Verify MEGA ticket close gates — ensures all sub-tickets Done, PRs merged, live verification passed.

The file

Show its rules and text

We split this skill into the rules it gives an agent

13 rules. About 0 are common, 8 are distinctive, and 5 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈8

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 5

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Run all 5 MEGA Close Gates on a ticket (or audit all recently-closed MEGAs).

    line 11
  2. Very specific:

    If $ARGUMENTS is --all-done-7d: Audit mode — scan all MEGAs marked Done in last 7 days

    line 16
  3. Very specific:

    If $ARGUMENTS is CAB-XXXX: Single mode — verify one MEGA ticket

    line 17
  4. Distinctive:

    If no children → report: "CAB-XXXX is not a MEGA (no children).

    line 32
  5. Distinctive:

    Run gates on each.

    line 44
  6. Very specific:

    Check that at least one comment contains PR # or PR [#.

    line 59
  7. Distinctive:

    Pass: All Done children have PR evidence

    line 61
  8. Very specific:

    Check each child's state.name:

    line 66
  9. Very specific:

    Pass: ALL children have state.name == "Done" or state.name == "Canceled"

    line 67
  10. Distinctive:

    Cross-component: verify all affected endpoints

    line 81
  11. Distinctive:

    If automated verification is not possible, prompt the user: "Gate 3 requires manual verification.

    line 83
  12. Distinctive:

    Check no child is stuck in non-terminal state:

    line 88
  13. Distinctive:

    Keep parent MEGA In Progress (do NOT mark Done)

    line 124

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Verify MEGA — $ARGUMENTS

Overview

Run all 5 MEGA Close Gates on a ticket (or audit all recently-closed MEGAs). Prevents false Dones on multi-phase tickets.

Mode Detection

  • If $ARGUMENTS is --all-done-7d: Audit mode — scan all MEGAs marked Done in last 7 days
  • If $ARGUMENTS is CAB-XXXX: Single mode — verify one MEGA ticket

Step 1: Fetch MEGA Data

Single Mode

linear.get_issue("CAB-XXXX", includeRelations=true)

Extract:

  • title — verify contains [MEGA] or has children
  • children.nodes[] — list of sub-tickets with id, identifier, state.name, estimate
  • estimate — total points

If no children → report: "CAB-XXXX is not a MEGA (no children). Use standard close flow."

Audit Mode

linear.list_issues(
  teamId: "<LINEAR_TEAM_ID>",
  filter: { state: { name: { eq: "Done" } }, updatedAt: { gte: "<7 days ago>" } }
)

Filter results to only MEGAs (title contains [MEGA] or has children). Run gates on each.

Step 2: Gate 0 — Decomposition Invariant

  • Check: children.nodes.length > 0
  • Pass: MEGA has sub-tickets
  • Fail: "MEGA has no sub-tickets. Run /decompose CAB-XXXX first."

Step 3: Gate 1 — Per-Phase PR Evidence

For each child ticket in Done state:

linear.list_comments(issueId: "<child_id>")

Check that at least one comment contains PR # or PR [#.

  • Pass: All Done children have PR evidence
  • Fail: List children missing PR references: "CAB-XXXX: no PR evidence found"

Step 4: Gate 2 — All Sub-Tickets Done

Check each child's state.name:

  • Pass: ALL children have state.name == "Done" or state.name == "Canceled"
  • Fail: List non-Done children with their current state:
    CAB-1351: In Progress
    CAB-1352: Todo

Step 5: Gate 3 — Live Verification

Based on the MEGA's component scope:

  • API changes: curl -s https://api.gostoa.dev/health → expect 200
  • UI/Portal: curl -s https://console.gostoa.dev or curl -s https://portal.gostoa.dev → expect 200
  • Gateway: curl -s https://mcp.gostoa.dev/health → expect 200
  • Docs: curl -s https://docs.gostoa.dev → expect 200
  • Cross-component: verify all affected endpoints

If automated verification is not possible, prompt the user: "Gate 3 requires manual verification. Please confirm the target is working: [endpoint/site]"

Step 6: Gate 4 — No Stale Children

Check no child is stuck in non-terminal state:

  • Pass: Zero children in "In Progress", "Todo", or "Blocked"
  • Fail: List stale children (these should be either Done or explicitly descoped)

Step 7: Render Verdict

Format

## MEGA Verification: CAB-XXXX — [Title]

| # | Gate | Status | Detail |
|---|------|--------|--------|
| 0 | Decomposition | ✅/❌ | N sub-tickets |
| 1 | PR Evidence | ✅/❌ | N/M children have PRs |
| 2 | All Done | ✅/❌ | N/M children Done |
| 3 | Live Verification | ✅/❌ | endpoint status |
| 4 | No Stale Children | ✅/❌ | N stale |

**Verdict**: Go / Fix

If Go (all gates pass)

  1. Mark parent MEGA as Done on Linear:
    linear.update_issue(id, state="Done")
  2. Post comprehensive completion comment:
    linear.create_comment(issueId, body="MEGA completed — all 5 gates passed.\n\nSub-tickets:\n- CAB-1350: Done (PR #578)\n- CAB-1351: Done (PR #582)\n- CAB-1352: Done (PR #590)\n\nLive verification: ✅ [endpoint] responding")
  3. Log: MEGA-GATE-PASS | task=CAB-XXXX gates=5/5

If Fix (any gate fails)

  1. Keep parent MEGA In Progress (do NOT mark Done)
  2. Post audit comment listing gaps:
    linear.create_comment(issueId, body="MEGA verification failed — X/5 gates passed.\n\nFailing gates:\n- Gate 1: CAB-1352 missing PR evidence\n- Gate 2: CAB-1352 still In Progress\n\nAction required: complete remaining work before closing.")
  3. Log: MEGA-GATE-FAIL | task=CAB-XXXX missing=<gate_numbers>
  4. List remediation steps for each failing gate

Audit Mode Output

## Weekly MEGA Audit — [date]

| MEGA | Points | Gates | Verdict | Action |
|------|--------|-------|---------|--------|
| CAB-1470 | 21 pts | 3/5 | Fix | Gate 1,2 failing |
| CAB-1490 | 13 pts | 5/5 | Go | Already correct |

Reopened: CAB-1470 (moved back to In Progress)

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.