setup-ticketing
Skill1 starsChanged 32 days agoMITView on GitHub
Install a lightweight, AI-first, in-repo ticket system into the
In this repository56 other files it gives its agents.
- autopilotSkill
- before-afterSkill
- bigpictureSkill
- blast-radiusSkill
- brainstormSkill
- compoundSkill
- confidence-checkSkill
- contextualizeSkill
- create-ship-skillSkill
- create-skillSkill
- curateSkill
- debugSkill
- deep-researchSkill
- devtoolsSkill
- distillSkill
- done-checkSkill
- doneSkill
- double-checkSkill
- envsetSkill
- explainSkill
- fortifySkill
- goSkill
- handoffSkill
- harness-auditSkill
- interviewSkill
- matrixSkill
- mockupSkill
- mutateSkill
- notesSkill
- postmortemSkill
- presentSkill
- qaSkill
- readableSkill
- recommendSkill
- reportSkill
- retireSkill
- review-planSkill
- scoutSkill
- sec-reviewSkill
- setup-worktreesSkill
- spec-checkSkill
- specSkill
- stepsSkill
- successionSkill
- super-reviewSkill
- tangentSkill
- tddSkill
- tend-memorySkill
- testSkill
- ticketSkill
- tldrSkill
- triageSkill
- waitSkill
- workSkill
- worth-itSkill
- wrap-upSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 37
Read the repo's CLAUDE.md first.
- Distinctive: line 90
Follow the project's commit convention (for a small solo repo, often commit+push directly).
≈ “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
Install the ticket system into the project in scope, then get out of the way. The system is the deliverable; this skill should never be needed twice in the same repo.
Why this shape (context for judgment calls, not dogma): it assumes you
interface with tickets through Claude Code — so the format optimizes for
model reads/writes over human browsing. One file per ticket keeps a ticket
one cheap targeted read. No index file, ever — in practice a
BACKLOG.md index bloats into inline prose walls and drifts from reality;
ls+grep over frontmatter answers every board question with zero drift.
Status flips in place (no moves to done/) so paths and IDs stay stable
forever.
0. Guard
- Explicit request only; never proactive.
- Read the repo's CLAUDE.md first. If it already declares a tracker — Jira, another system, or a previous tickets/ install — surface that and stop. One repo, one tracker; migrating is a different conversation.
- A
tickets/dir that already exists without a CLAUDE.md block is a half-install: say so and offer to finish (write the block), not re-init.
1. Derive the prefix
Uppercase initials of the hyphenated repo name: acme-budget → AB,
dnd-ai-vault → DAV. $ARGUMENTS overrides. State the prefix in the
final report — it's the one thing worth the user noticing. IDs are PREFIX-NNN,
zero-padded to 3, sequential, never reused — a dropped ticket's number is
retired with it.
2. Create the structure
tickets/ at repo root (+ .gitkeep only if no tickets are filed in the
same session). Nothing else — no README, no template file, no index. The
conventions live in one place (CLAUDE.md block below); duplicating them into
tickets/ would just create a second copy to drift.
3. Write the CLAUDE.md block
Append to the project's CLAUDE.md (create the file if the repo has none), adapting the prefix. This block is the whole daily-use system:
## Tickets (in-repo tracker)
This repo's tracker is `tickets/` — one markdown file per ticket,
`tickets/AB-NNN-short-slug.md`. No external tracker, no index file: `ls`+`grep`
is the board. Skills that file or check tickets (postmortem,
done-check, handoff) file HERE.
---
id: AB-NNN # PREFIX + 3-digit seq; next = highest existing + 1; never reuse
title: One-line summary
status: open # open | doing | done | dropped — flip IN PLACE, never move the file
priority: p2 # p1 now-ish | p2 normal | p3 someday
created: YYYY-MM-DD
blocked_by: [AB-004] # optional — omit unless real
tags: [artifact] # optional — omit unless useful
---
Body: freeform, as short as the idea allows — why it matters, rough
acceptance shape, links. On close, append one line:
"Done YYYY-MM-DD: <what shipped, sha>" or "Dropped: <why>".
Live board: `grep -H '^status:' tickets/*.md | grep -v -e done -e dropped`
File tickets for: decided follow-ups, parked ideas worth keeping, scoped
chunks of future work. Not for in-flight session state (that's the
session/handoff), not for someday-maybes not worth a file.4. Land it
Follow the project's commit convention (for a small solo repo, often commit+push directly). If tickets are being filed in the same breath — the usual dogfood case — file them first, one commit for the lot.
5. Report
One line: prefix, path, board command, and (if any) tickets filed. The right follow-up isn't a skill health-check — it's simply filing the first ticket.
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.

