triage Skill (mattpocock/skills) for Claude Code: What It Does and How to Install It

Helps a maintainer move issues and outside pull requests through a fixed set of labels and turn them into briefs an AI agent can pick up.

Installs: 1M; on skills.sh
Stars: 309.2K; on GitHub
Files: 3; besides SKILL.md
Updated: Nov 2026; last repo commit

triage makes Claude work through your issue tracker with you. You type /triage and say what you want, such as "show me anything that needs my attention" or "let's look at #42". It reads the issue, checks your code, reproduces bugs, recommends a label and waits for your call.

Every issue gets one category (bug or enhancement) and one state (needs-triage, needs-info, ready-for-agent, ready-for-human or wontfix). It posts comments and changes labels on your tracker, and every comment starts with an AI-generated disclaimer. Run setup-matt-pocock-skills first so it knows your tracker and label names. It runs no bundled scripts; it uses your gh or glab tool.

Good for

Open-source maintainers with an issue backlog
Teams that want issues turned into briefs an AI agent can implement
Developers already using the other mattpocock engineering skills

Install triage

Claude Code

Copy this prompt.

Install the triage skill in Claude Code. Run this command:

claude plugin install mattpocock-skills@claude-plugins-official

Then check that it loaded and tell me how to use it.

2
Paste it into Claude Code. It shows the change and asks before running it.

Pick one route, not both, or you get every skill twice. The plugin installs the whole set and updates itself. The skills CLI route must name grilling, domain-modeling, setup-matt-pocock-skills too, because triage hands work to them. Claude Code names plugin skills /plugin-name:skill-name, so after the plugin install the full name is /mattpocock-skills:triage; the README lists /triage, which is not confirmed for plugin installs. The README also asks you to run /setup-matt-pocock-skills once per repo.

From the project README. Source.

What's in the skill

When it loads

Move issues and external PRs through a state machine of triage roles, categorise, verify, grill if needed, and write agent-ready briefs.

Sections in SKILL.md

Reference docs
Roles
Invocation
Show what needs attention
Triage a specific issue or PR
Quick state override
Needs-info template
Triage Notes
Resuming a previous session

Key facts

Author: Matt Pocock (mattpocock/skills)
License: MIT
Installs on skills.sh: 1,025,908 (Nov 2, 2026)
GitHub stars: 309,237 (whole repo, Nov 2, 2026)
Last commit: Nov 1, 2026 (repo)
Files it bundles: SKILL.md, AGENT-BRIEF.md, OUT-OF-SCOPE.md, agents/openai.yaml
Starts: Only when you call it

What it includes

Lists unlabeled, needs-triage and updated needs-info items, oldest first
Network · Show what needs attention
Gathers context, checks for duplicates and past rejections, then recommends a label
Files · Triage one issue or PR
Reproduces the bug or runs the PR's tests and reports what happened
Runs code · Verify the claim
Calls grilling and domain-modeling to flesh out a vague request
Write · Grill if needed
Posts an agent brief, needs-info notes or closes as wontfix
Write · Apply the outcome
"Move #42 to ready-for-agent" applies the label after you confirm
Write · Quick state override
Guides for agent briefs and a .out-of-scope/ list of rejected requests
Write · AGENT-BRIEF.md and OUT-OF-SCOPE.md

From the official docs

Permissions

Medium risk

What it can do

Read your data
Create and change things
Delete things
Read-only modeHow to turn it on

Watch out for

Runs code from outside PRs
To verify a PR it checks it out and runs the relevant tests or commands. Do that only in an environment you are happy to risk.
Changes your tracker
It posts comments, changes labels and closes issues. Every comment it posts starts with an AI-generated disclaimer.
Two audits warn
Gen Agent Trust Hub and Snyk warn on skills.sh; Socket passes. skills.sh

Technical details

Turning on read-only	Ask it to show what needs attention or to recommend a label. It waits for your direction before posting or changing anything.
Talks to	Your issue tracker (GitHub or GitLab, through gh or glab)
Sign-in and keys	Uses whatever gh or glab is already signed in to. No separate key.
Scopes	Reads issues, pull requests and comments through gh or glab Posts comments, applies labels and closes issues as wontfix Can check out a PR and run its tests or commands to verify it Writes .out-of-scope/ notes and, during grilling, GLOSSARY.md or ADRs

Questions

What does the triage skill do?

It runs issues and external PRs through a small state machine: two categories and five states. For each, it gathers context, verifies the claim, recommends a label and posts a brief or notes.

How do I install it?

Use the install box at the top of this page. It installs triage with grilling, domain-modeling and setup-matt-pocock-skills, which triage depends on.

Do I need to run the setup skill first?

Yes. triage needs your label names and tracker details. If the mapping is missing, it tells you to run /setup-matt-pocock-skills.

Can it change or close my issues?

Yes. It posts comments, applies labels and can close issues as wontfix. It confirms with you first and starts each comment with an AI-generated disclaimer.

Does it need an API key?

No separate key. It uses whatever your issue tracker command line (such as gh) is already signed in to.

Is it safe?

It can write to your tracker, so use a repo you control and read each action before you confirm. skills.sh audits: Gen Agent Trust Hub warn, Socket pass, Snyk warn.

Information

updated Nov 2, 2026

Repository: mattpocock/skills
Folder: skills/engineering/triage

mattpocock/skills skills.sh

Sources

Docs, issue trackers and discussions this page is built from.

Show all 6

mattpocock/skills README · github.com
skills CLI commands · github.com
Claude Code skills · code.claude.com
triage SKILL.md · github.com
triage on skills.sh · skills.sh
mattpocock/skills · github.com