epic-cleanup
Skill1 starsChanged 6 months agoMITView on GitHub
Audit all active epics for ownership gaps, stale work, and closure candidates. Produces an actionable checklist: assign DRIs, close completed/dead epics, and make decisions on ambiguous work. Organized by person then by action type. Default output is conversational; optionally save to vault or file.
In this repository30 other files it gives its agents.
- setupSkill
- enrich-prsSkill
- epic-overviewSkill
- oncall-reviewSkill
- sprint-managementSkill
- sprint-statusSkill
- team-activitySkill
- team-data-syncSkill
- direct-report-prepSkill
- interview-analysisSkill
- interview-prepSkill
- peer-prepSkill
- project-statusSkill
- team-analysisSkill
- today-prioritiesSkill
- weekly-snippetSkill
- bus-factorSkill
- contribution-velocitySkill
- nine-boxSkill
- project-people-mapSkill
- review-cardsSkill
- review-cyclesSkill
- review-networkSkill
- team-health-reportSkill
- task-rolloverSkill
- context-checkSkill
- prep-my-daySkill
- initiative-breakdownSkill
- product-briefSkill
- technical-deep-diveSkill
The file
We split this skill into the rules it gives an agent
4 rules. About 0 are common, 3 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈3
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 30
Default: conversational (printed to screen).
- Very specific: line 99
If user asks to save to vault: Use - [ ] task checkboxes so items are trackable.
- Distinctive: line 114
Be direct about recommendations: Don't just flag "unassigned" — suggest who should own it based on their existing work.
- Distinctive: line 118
Call out load imbalance: If one person has 8 epics and another has 0, say so.
≈ “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.
Files
Epic Cleanup
Audit all active epics and initiatives for ownership gaps, staleness, and hygiene issues. Produces a person-by-person ownership map plus actionable task lists for assigning, closing, and deciding on epics.
When to use
- "Epic cleanup", "audit our epics", "who owns what"
- "Which epics need a DRI", "what should we close"
- Sprint planning: "clean up the board before we start"
- Quarterly: "are all our epics still relevant"
Prerequisites
- GitHub CLI (
gh) authenticated with org access - Team repo with issues labeled
epic,initiative, orbatch - Team roster (
team.json) for member identification
Output
Default: conversational (printed to screen). User can ask to save it as a vault note with task checkboxes for tracking.
Process
Step 1: Gather all open epics and parent issues
Same as epic-overview: query for all open issues that are epics, initiatives, batches, or have sub-issues.
Step 2: Build ownership map
For each team member, list every epic/initiative where they are assigned (DRI). Include:
- Issue number and title (linked)
- Progress (sub-issues completed / total)
- Activity status: ✅ Active, ⚠️ Stale (14+ days), ❓ No signal (30+ days)
Step 3: Identify action items
Categorize into:
Assign immediately — Epics with active work happening (sprint goals, recent PRs) but no DRI assigned.
Assign or close (nearly done) — Epics at 80%+ completion with no DRI to close out the remaining items.
Probably close — Epics that are:
- 100% complete but still open
- Superseded by newer work
- Stale for 60+ days with no indication of resuming
- From a previous fiscal year with no recent relevance
Needs a decision — Epics where it's unclear if the work is still planned, deprioritized, or abandoned. Surface these for the manager to make a call.
Unassigned parent issues — Issues acting as epics (have sub-issues) but no one is assigned. Need a DRI decision.
Step 4: Write the output
Person-by-person section:
### @login (Name)
- [#N Title](url) — X/Y (Z%) status_emoji description
- [#N Title](url) — X/Y (Z%) status_emoji descriptionAction items section (when saving to vault, use checkboxes):
## Suggested Actions
**Assign immediately:**
- [ ] [#N](url) Assign to [[Person]] (reason)
**Assign or close:**
- [ ] [#N](url) Title (X%) — who finishes this?
**Probably close:**
- [ ] [#N](url) Title (reason it's dead)
**Needs a decision:**
- [ ] [#N](url) Title — still planned or deprioritized?
**Unassigned:**
- [ ] [#N](url) Title (X/Y) — suggested owner?Step 5: Output
Default: Print to screen as markdown (no checkboxes, just bullets with recommendations).
If user asks to save to vault: Use - [ ] task checkboxes so items are trackable. Include wikilinks to People files for assignee suggestions. Link all issue numbers.
Staleness thresholds
| Days since last activity | Signal |
| 0-14 | ✅ Active |
| 14-30 | ⚠️ Getting stale |
| 30-60 | ❓ No signal, needs check-in |
| 60+ | 🔴 Likely dead or deprioritized |
Activity = any of: comment posted, sub-issue closed, PR merged referencing the epic, label change, assignee change.
Content guidelines
- Be direct about recommendations. Don't just flag "unassigned" — suggest who should own it based on their existing work.
- Group actions by effort level. "Assign immediately" is 30 seconds of work. "Needs a decision" requires a conversation.
- Use wikilinks when saving to vault. Link People files so the user can navigate to context.
- Link every issue. Full URLs so items are one click away.
- Call out load imbalance. If one person has 8 epics and another has 0, say so.
- Don't mix cleanup with status. This isn't about "how is work going" — it's "is our board accurate and owned."
Relationship to other skills
- epic-overview: Overview is the "what exists" reference. Cleanup is the "what needs fixing" action list.
- project-status: Status reports on weekly progress. Cleanup is periodic hygiene (monthly or quarterly).
- sprint-management: Sprint management moves items between sprints. Cleanup ensures the backlog of epics is healthy.
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
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.

