Github triage

Skill OutlineDriven/odin-claude-plugin/skills/github-triage

Outline-Driven Development for Claude Code - 46 agents, 25+ skills, diagram-first methodology, AST-based editing, atomic commits.

Install

npx -y skills add OutlineDriven/odin-claude-plugin --skill github-triage

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

What its author says it does

Copied from the file, not written here

Triage GitHub issues. Use when the user wants to triage incoming issues, prepare issues for an autonomous agent, or move an issue between workflow states.

SKILL.md

4.3 KB, 843 tokens by cl100k_base, as published. Nobody here has run it

State-machine triage over GitHub issues. Every comment posted carries an AI disclaimer. Label names are configurable; state semantics are not.

AI disclaimer (mandatory)

Every comment or new issue body posted during triage MUST begin with:

> *This was generated by AI during triage.*

No exceptions.

Label map (configure per project)

Bind ODIN's abstract state tokens to your repo's actual label names. Edit this block before first use; defaults match Matt Pocock's upstream convention.

category:
  bug:        "bug"             # something is broken
  enhancement:"enhancement"     # new feature or improvement
state:
  triage_pending:   "needs-triage"     # maintainer must evaluate
  awaiting_info:    "needs-info"       # waiting on reporter
  ready_for_agent:  "ready-for-agent"  # fully specified, autonomous-agent-suitable
  ready_for_human:  "ready-for-human"  # requires human judgment
  wont_fix:         "wontfix"          # closed, not actioned

Body uses the abstract tokens (triage_pending, ready_for_agent, etc.). Replace with your repo's labels at apply time.

Invariants

Every issue carries exactly one category label and one state label.
Conflicting state labels = halt and ask the maintainer which is correct.
Maintainer can override any state directly; flag unusual transitions.

State machine

From	To	Trigger	Effect
unlabeled	triage_pending	skill (first look)	applies label after presenting recommendation
unlabeled	ready_for_agent	maintainer	post agent brief, apply label
unlabeled	ready_for_human	maintainer	post task summary, apply label
unlabeled	wont_fix	maintainer	comment + close
triage_pending	awaiting_info	maintainer	post triage notes capturing progress + reporter questions
triage_pending	ready_for_agent	maintainer	grilling complete; agent brief; apply label
triage_pending	ready_for_human	maintainer	grilling complete; task summary; apply label
triage_pending	wont_fix	maintainer	comment + close
awaiting_info	triage_pending	skill (reply seen)	reporter replied; surface for re-evaluation

Workflows

Overview ("what needs my attention")

Group open issues into three buckets and display oldest-first:

Unlabeled: never triaged.
triage_pending: needs evaluation or continuation.
awaiting_info with new activity since the last triage-notes comment.

Triage a specific issue

Dispatch Explore agent to read full issue + all comments + related code paths.
Present category and state recommendations with reasoning.
For bugs: attempt reproduction. Run repo tests, trace logic.
If underspecified, do a focused domain-model pass: map entities and constraints before re-categorizing.
Apply outcome per state.

`awaiting_info` comment template

The Triage Notes template (established-so-far / need-from-reporter) lives in references/awaiting-info-template.md — read it when transitioning an issue to awaiting_info; other transitions (ready_for_agent, ready_for_human, wont_fix) don't need it.

Parallel examples

TypeScript repo: label-map binds ready_for_agent → "agent-ready", awaiting_info → "blocked-on-reporter". State machine and prose unchanged.

Rust repo: label-map binds triage_pending → "S-needs-triage", wont_fix → "C-wont-fix" per Rust-project conventions. State machine and prose unchanged.

Gives 0 of the 12 instructions most debug triage skills give in 843 tokens

Counted across 839 of the 1,149 authors here whose files we hold, read 2026-08-06

investigate root cause before proposing any fixin 102 of 839, across 65 files
read error messages completelyin 90 of 839, across 48 files
create a failing test case before fixingin 84 of 839, across 44 files
reproduce the issue consistentlyin 82 of 839, across 40 files
change one variable at a timein 82 of 839, across 42 files
check recent changesin 74 of 839, across 35 files
write the regression test before fixingin 74 of 839, across 36 files
fix the root cause not the symptomin 60 of 839, across 43 files
implement a single fix at a timein 59 of 839, across 20 files
trace data flow backward to the sourcein 50 of 839, across 20 files
remove all debug instrumentationin 49 of 839, across 13 files
form a single hypothesisin 48 of 839, across 18 files

Said here and by no other author read

Flag unusual state transitions
Read the full issue and related code paths
Map entities and constraints if the issue is underspecified
Read the template file before transitioning to awaiting_info
Post task summaries for ready_for_human transitions

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.