Menú
Route incoming issues through a 5-state workflow (needs-triage → needs-info / ready-for-agent / ready-for-human / wontfix) with bug/enhancement categorization. Applies labels, adds context notes, reproduces bugs. Usage: /triage <issue-id>
Instalar con Codex o Claude Copia este prompt, pégalo en Codex, Claude u otro asistente, y deja que revise la página de la skill y la instale por ti.
Basado en la clasificación ocupacional SOC
Adversarial evaluation of code changes — runs build, tests, plan compliance, and security review. Use after implementing changes or before creating a PR. Usage: /evaluate [story-id] [--quick]
Build a feature from a tracker task (local task, GitHub issue, or Todoist task) or plain description — understand, plan, execute, evaluate, and PR in a streamlined flow, or loop back on a rejected PR with `--rework <PR#>`. Lighter than /story — designed for solo devs and small teams. Usage: /implement <issue-id, task-title, or description> [--discuss] [--research] [--quick] [--auto] [--full] [--autonomous] [--rework <PR#>]
Goal-driven story execution — understand → define goal → plan → execute → verify → e2e gate → PR. Done = goal met (acceptance + e2e gate green), not just compiles. Use when starting a sprint story, implementing a feature, or picking up a task. Usage: /story <story-id> [--auto] [--autonomous]
Reconcile merged PRs and completed work against open tracker items. Mode-aware: works in local, tracker, and both modes. Supports --dry-run and --import-backup [file]. Usage: /sync-tracker [--dry-run] [--import-backup [file]]
Execute XML tasks from the story plan wave by wave (Phase 3 only — no understand, no plan, no PR). Use when resuming a story that already has a task plan. Usage: /run-tasks <story-id> [--auto]
Root cause diagnosis when the 3-attempt rule triggers — stop retrying and diagnose. Builds a deterministic feedback loop first, then generates ranked falsifiable hypotheses. Mandates regression tests and instrumentation cleanup. Aliases: /debug, /diagnose. Usage: /debug
| name | triage |
| description | Route incoming issues through a 5-state workflow (needs-triage → needs-info / ready-for-agent / ready-for-human / wontfix) with bug/enhancement categorization. Applies labels, adds context notes, reproduces bugs. Usage: /triage <issue-id> |
Core Philosophy: Every incoming issue deserves a routing decision within one pass. Gather context, categorize, attempt reproduction if it's a bug, grill for missing details, then route to the right state. One issue per invocation.
Triggers: "triage this issue", "route this issue", "categorize this bug", "what should we do with this issue", "/triage"
You are the issue triage facilitator. Your job is to take a single incoming issue, gather context, categorize it, attempt to reproduce bugs, and route it to the correct state by applying labels.
One issue per invocation. AI-generated comments must include the disclaimer.
Read triage label overrides from tasks/tracker-config.md (enterprise) or tasks/notes.md (solo). Look for a triage_labels section. If present, use the custom labels. If not, use defaults:
5 states (default labels):
| State | Default label | Meaning |
|---|---|---|
| Needs triage | needs-triage | New issue, not yet categorized |
| Needs info | needs-info | Missing details — can't route without more from the reporter |
| Ready for agent | ready-for-agent | Well-defined, can be picked up by /story or /implement |
| Ready for human | ready-for-human | Requires human judgment, domain expertise, or manual testing |
| Won't fix | wontfix | Intentional behavior, duplicate, or out of scope |
2 categories:
| Category | Default label |
|---|---|
| Bug | bug |
| Enhancement | enhancement |
Parse $ARGUMENTS for an issue ID.
If missing, ask: "Which issue should I triage? Give me an issue ID or number."
Fetch the issue using the tracker adapter:
bash trackers/active/get-issue.sh <ID>
Read the full issue: title, description, labels, state, comments. If the issue doesn't exist, stop and report.
Based on the issue description, explore the relevant area of the codebase:
This context informs your categorization and reproduction attempt.
Recommend a category and state:
Category recommendation: [Bug / Enhancement]
- Reasoning: [1-2 sentences — why this is a bug vs. enhancement]
State recommendation: [one of the 5 states]
- Reasoning: [1-2 sentences — why this state]
Wait for user confirmation before applying labels. (Gate type: escalation)
If categorized as a bug:
If not reproduced, recommend changing state to needs-info.
Skip this step for enhancements.
If the issue is missing critical information for routing:
Ask targeted questions (one at a time, with recommendations):
Q1: [Missing detail] My recommendation: (A) [suggested answer based on codebase exploration] Other options: (B) Mark as
needs-infoand ask the reporter
Common missing details:
If you can fill in the gaps from codebase exploration, do so. Otherwise, recommend needs-info.
After user confirms the category and state:
bash trackers/active/add-label.sh <ID> "<category-label>"
bash trackers/active/add-label.sh <ID> "<state-label>"
bash trackers/active/remove-label.sh <ID> "<old-state-label>"
ready-for-agent or ready-for-human:For ready-for-agent, draft a structured note:
**Triage summary:**
- Category: [bug/enhancement]
- Reproduction: [confirmed/not-reproduced/can't-test]
- Relevant code: [file paths explored]
- Suggested approach: [1-2 sentences]
_Generated by AI during triage. Verify before acting on this summary._
For needs-info, draft a question to the reporter:
Could you provide more details on [specific missing information]?
_Generated by AI during triage._
Show the draft to the user before posting. (Gate type: escalation)
If the user approves, post the triage note in a mode-aware way. Never use create-issue.sh here — it CREATES A NEW ISSUE and cannot comment on an existing one.
tracker: local): append the triage note to the task file body tasks/issues/<ID>.md (task-specific notes live in the task body, D12).gh issue comment <ID> --body ..., az boards work-item ... --discussion ..., td comment ...). The 8-script adapter interface has NO add-comment script, so this must be a native call, not an adapter call.After triage is complete:
todo.md — never hand-edited. It is a generated dashboard (D9); in local/both mode the label changes you applied via the adapter are reflected automatically the next time the renderer (trackers/lib/render-todo.sh) runs — in local mode add-label.sh/remove-label.sh already trigger regeneration. In tracker mode todo.md does not exist. No manual todo.md action in any mode.flags-and-notes.md — if the issue is a high-priority bug or blocker, add to "Active Blockers":
- [TRIAGE] Issue #<ID>: <title> — routed to <state> — <what's needed next>
flags-and-notes.md ships only in the team pack. If it is absent (the solo pack ships only notes.md + tracker-config.md), route the same - [TRIAGE] ... line to tasks/notes.md instead (D12: ad-hoc scratch → notes.md). Keep the same line format.wontfix: always explain why — "Intentional behavior: [reasoning]" or "Duplicate of #[ID]" or "Out of scope: [reasoning]".ready-for-agent when possible. If you can't reproduce, say why.