triage

star 0

Triage Linear issues, Projects, PRDs, bugs, and feature requests through the repo's Linear state machine. Use when the user wants to create, classify, clarify, label, unblock, close, or prepare Linear work for human or agent execution.

cill-i-am By cill-i-am schedule Updated 6/11/2026

name: triage description: Triage Linear issues, Projects, PRDs, bugs, and feature requests through the repo's Linear state machine. Use when the user wants to create, classify, clarify, label, unblock, close, or prepare Linear work for human or agent execution.

Triage

Move Linear work through the state machine defined in docs/agents/triage-states.md. Triage is about making work truthful: either ready for an AFK worker, waiting for information, blocked, ready for a human, or explicitly out of scope.

Read First

  • docs/agents/linear-workflow.md
  • docs/agents/triage-states.md
  • docs/agents/domain.md
  • relevant Linear issue/Project/PRD/comments
  • relevant source and architecture docs for the area

Use the Linear skill/app for reads and writes. If Linear tools are unavailable, stop and ask the user to connect Linear.

Show What Needs Attention

When asked what needs attention, query Linear and group oldest first:

  1. Untriaged or unlabeled work.
  2. needs-triage.
  3. needs-info with new reporter or maintainer activity.
  4. blocked issues whose blockers changed.
  5. in-review issues missing PR, CI, or orchestrator evidence.

Show counts and one-line summaries. Let the user pick unless they asked for a specific issue.

Triage A Specific Issue

  1. Gather context. Read the full issue, comments, labels/status, parent Project/PRD, blockers, assignee, and related PRs. Check relevant source or docs rather than trusting stale issue prose.
  2. Classify. Recommend category and state: bug, enhancement, chore, spike; and one of the states in docs/agents/triage-states.md.
  3. Reproduce bugs when possible. Trace the likely code path and run focused commands when cheap. If reproduction needs credentials, provider state, or an Alchemy stage, state that clearly.
  4. Clarify if needed. Use grill-with-docs for real domain ambiguity. Capture resolved decisions in Linear comments or the parent PRD.
  5. Prepare for execution. If ready for an agent, ensure the issue has acceptance criteria, blockers, verification expectations, out-of-scope boundaries, and AFK/HITL classification.
  6. Apply Linear updates. Update status/labels/relations and add a concise comment explaining what changed.

Ready-For-Agent Checklist

Before moving an issue to ready-for-agent, verify:

  • parent Project/PRD is linked or intentionally absent
  • acceptance criteria are concrete and testable
  • blockers are represented as Linear relations
  • no unresolved HITL decision is hidden in prose
  • out-of-scope boundaries are explicit
  • verification expectations are realistic
  • likely review stack and risk tier are noted

Wontfix / Out Of Scope

For rejected enhancements, record the reasoning durably:

  • Prefer a Linear Project/issue comment or linked document if this is purely product scope.
  • Use a repo doc only when the decision affects future architecture or agent behavior.

The reason must be durable. "Not now" is usually needs-info or deferred work, not wontfix.

Comments

Every triage comment should include:

  • current recommendation
  • evidence or code/docs checked
  • open questions or blockers
  • next state and why
Install via CLI
npx skills add https://github.com/cill-i-am/ceird --skill triage
Repository Details
star Stars 0
call_split Forks 0
navigation Branch main
article Path SKILL.md
More from Creator