triage — for Claude Code triage, community, for Claude Code, ide skills, execute, validate, feature, phase:1, phase:2, phase:3

v1.0.0

Об этом навыке

Подходящий сценарий: Ideal for AI agents that need after /audit produces findings that need sequencing. Локализованное описание: # Hali Triage Skill Purpose Transform a noisy issue list or audit findings into a clean, prioritized, actionable queue. This AI agent skill supports Claude Code, Cursor, and Windsurf workflows.

Возможности

After /audit produces findings that need sequencing
When the GitHub issue list has grown noisy and needs a cull pass
At the start of a new workstream to establish priority order
When multiple issues overlap and you suspect duplicates
You need to inspect the codebase first → /audit

# Core Topics

irvinesunday irvinesunday
[0]
[0]
Updated: 4/24/2026

Killer-Skills Review

Decision support comes first. Repository text comes second.

Reference-Only Page Review Score: 10/11

This page remains useful for teams, but Killer-Skills treats it as reference material instead of a primary organic landing page.

Original recommendation layer Concrete use-case guidance Explicit limitations and caution Quality floor passed for review
Review Score
10/11
Quality Score
70
Canonical Locale
en
Detected Body Locale
en

Подходящий сценарий: Ideal for AI agents that need after /audit produces findings that need sequencing. Локализованное описание: # Hali Triage Skill Purpose Transform a noisy issue list or audit findings into a clean, prioritized, actionable queue. This AI agent skill supports Claude Code, Cursor, and Windsurf workflows.

Зачем использовать этот навык

Рекомендация: triage helps agents after /audit produces findings that need sequencing. Hali Triage Skill Purpose Transform a noisy issue list or audit findings into a clean, prioritized, actionable queue. This AI agent

Подходит лучше всего

Подходящий сценарий: Ideal for AI agents that need after /audit produces findings that need sequencing.

Реализуемые кейсы использования for triage

Сценарий использования: Applying After /audit produces findings that need sequencing
Сценарий использования: Applying When the GitHub issue list has grown noisy and needs a cull pass
Сценарий использования: Applying At the start of a new workstream to establish priority order

! Безопасность и ограничения

  • Ограничение: After /audit produces findings that need sequencing
  • Ограничение: When the GitHub issue list has grown noisy and needs a cull pass
  • Ограничение: You need to inspect the codebase first → /audit

Why this page is reference-only

  • - Current locale does not satisfy the locale-governance contract.

Source Boundary

The section below is imported from the upstream repository and should be treated as secondary evidence. Use the Killer-Skills review above as the primary layer for fit, risk, and installation decisions.

After The Review

Decide The Next Action Before You Keep Reading Repository Material

Killer-Skills should not stop at opening repository instructions. It should help you decide whether to install this skill, when to cross-check against trusted collections, and when to move into workflow rollout.

Labs Demo

Browser Sandbox Environment

⚡️ Ready to unleash?

Experience this Agent in a zero-setup browser environment powered by WebContainers. No installation required.

Boot Container Sandbox

FAQ & Installation Steps

These questions and steps mirror the structured data on this page for better search understanding.

? Frequently Asked Questions

What is triage?

Подходящий сценарий: Ideal for AI agents that need after /audit produces findings that need sequencing. Локализованное описание: # Hali Triage Skill Purpose Transform a noisy issue list or audit findings into a clean, prioritized, actionable queue. This AI agent skill supports Claude Code, Cursor, and Windsurf workflows.

How do I install triage?

Run the command: npx killer-skills add irvinesunday/Hali/triage. It works with Cursor, Windsurf, VS Code, Claude Code, and 19+ other IDEs.

What are the use cases for triage?

Key use cases include: Сценарий использования: Applying After /audit produces findings that need sequencing, Сценарий использования: Applying When the GitHub issue list has grown noisy and needs a cull pass, Сценарий использования: Applying At the start of a new workstream to establish priority order.

Which IDEs are compatible with triage?

This skill is compatible with Cursor, Windsurf, VS Code, Trae, Claude Code, OpenClaw, Aider, Codex, OpenCode, Goose, Cline, Roo Code, Kiro, Augment Code, Continue, GitHub Copilot, Sourcegraph Cody, and Amazon Q Developer. Use the Killer-Skills CLI for universal one-command installation.

Are there any limitations for triage?

Ограничение: After /audit produces findings that need sequencing. Ограничение: When the GitHub issue list has grown noisy and needs a cull pass. Ограничение: You need to inspect the codebase first → /audit.

How To Install

  1. 1. Open your terminal

    Open the terminal or command line in your project directory.

  2. 2. Run the install command

    Run: npx killer-skills add irvinesunday/Hali/triage. The CLI will automatically detect your IDE or AI agent and configure the skill.

  3. 3. Start using the skill

    The skill is now active. Your AI agent can use triage immediately in the current project.

! Reference-Only Mode

This page remains useful for installation and reference, but Killer-Skills no longer treats it as a primary indexable landing page. Read the review above before relying on the upstream repository instructions.

Upstream Repository Material

The section below is imported from the upstream repository and should be treated as secondary evidence. Use the Killer-Skills review above as the primary layer for fit, risk, and installation decisions.

Upstream Source

triage

# Hali Triage Skill Purpose Transform a noisy issue list or audit findings into a clean, prioritized, actionable queue. This AI agent skill supports Claude

SKILL.md
Readonly
Upstream Repository Material
The section below is imported from the upstream repository and should be treated as secondary evidence. Use the Killer-Skills review above as the primary layer for fit, risk, and installation decisions.
Supporting Evidence

Hali Triage Skill

Purpose

Transform a noisy issue list or audit findings into a clean, prioritized, actionable queue. Deduplicate aggressively. Recommend which issues to close, merge, defer, relabel, or implement — and in what order. Does not write code.

When to use

  • After /audit produces findings that need sequencing
  • When the GitHub issue list has grown noisy and needs a cull pass
  • At the start of a new workstream to establish priority order
  • When multiple issues overlap and you suspect duplicates

When NOT to use

  • You need to inspect the codebase first → /audit
  • You are ready to implement a specific issue → /execute
  • You are reviewing a PR → /validate
  • A single issue is already clearly scoped and approved → go directly to /execute

Required pre-read

CLAUDE.md
docs/arch/HALI_PHASE_BLOCKERS.md
docs/arch/HALI_PHASE_SCAFFOLD.md
docs/arch/label-taxonomy.md

If label-taxonomy.md does not exist on disk at the time of the triage run, fall back to the label conventions observed in the repo: feature:, fix:, chore:, docs:, phase:1, phase:2, phase:3, blocked:, wont-fix.

Inputs expected

InputRequiredNotes
SourceYesaudit-report / issue-list / both
ScopeRecommendedphase:1 / phase:2 / backend / mobile / all
Audit reportIf source=audit-reportOutput from /audit
Issue rangeNoGitHub issue numbers or label filter

Example: /triage source=issue-list scope=phase:1

Procedure

Step 1 — Gather raw input

If source includes issue-list:

bash
1gh issue list --repo irvinesunday/Hali --state open --limit 100 \ 2 --json number,title,labels,createdAt,updatedAt,body

For each issue, also fetch the current state (verify open/closed):

bash
1gh issue view <NUMBER> --repo irvinesunday/Hali --json state,title,labels

Never assume an issue is open without checking.

If source includes audit-report: read the supplied audit findings before proceeding.

Step 2 — Classify each issue

Apply exactly one classification per issue:

ClassificationMeaning
IMPLEMENTIn-scope, actionable, no known blockers — enter the implementation queue
DEFERValid but not yet time — dependencies unmet or out of current phase
CLOSESuperseded, already delivered, invalid, or scope-violating
MERGE-DUPLICATEFunctionally identical to another open issue — merge into canonical
RELABELContent is valid but mislabeled — update labels only
BLOCKEDDepends on an issue that must land first — flag blocker explicitly
NEEDS-HUMAN-DECISIONScope or priority requires explicit owner decision before action

Step 3 — Apply phase and out-of-scope filters

Cross-reference each IMPLEMENT candidate against CLAUDE.md Phase 1 out-of-scope list:

  • Comments / replies / threads → CLOSE as out-of-scope
  • Likes / reactions → CLOSE as out-of-scope
  • Messaging / chat → CLOSE as out-of-scope
  • Media uploads → CLOSE as out-of-scope
  • Admin portal (before Phase 2 gating) → DEFER
  • Microservices → CLOSE as doctrine violation

Flag any issue requesting a locked architectural decision change (H3 → geohash, PostgreSQL → other, etc.) as NEEDS-HUMAN-DECISION.

Step 4 — Deduplicate

For issues with overlapping scope:

  1. Identify the canonical issue (oldest with most complete description, or explicitly linked to a spec)
  2. Mark duplicates as MERGE-DUPLICATE referencing the canonical number
  3. Propose closing the duplicates with a comment pointing to the canonical

Do not preserve issue noise just because it exists.

Step 5 — Sequence the IMPLEMENT queue

Order IMPLEMENT issues by:

  1. Phase gate (Phase 1 before Phase 2 before Phase 3)
  2. Dependency order (blocking issues before dependent issues)
  3. Risk surface (auth/schema changes before feature additions)
  4. Value/stability ratio (low-risk high-value first)

Flag any cross-issue dependencies explicitly:

  • "Issue #N must land before #M because [reason]"

Step 6 — Detect phase gate violations

Check CLAUDE.md — Phase 2 full UI delivery must not begin until Phase 1 mobile app is in pilot state. Any Phase 2 UI issue in IMPLEMENT status must be flagged with:

PHASE_GATE: Verify Phase 1 mobile pilot state before scheduling.

Step 7 — Draft actions

For each CLOSE recommendation: draft a brief, factual close reason. For each MERGE-DUPLICATE: identify the canonical issue number and draft a close comment. For each RELABEL: specify current labels → proposed labels.

Do not execute any of these actions without explicit approval unless the /triage invocation explicitly includes --apply.

Output requirements

=== HALI TRIAGE REPORT ===

Source: [input type]
Scope: [declared scope]
Issues reviewed: [count]
Date: [ISO 8601]

Classification summary:
- IMPLEMENT:         [count]
- DEFER:             [count]
- CLOSE:             [count]
- MERGE-DUPLICATE:   [count]
- RELABEL:           [count]
- BLOCKED:           [count]
- NEEDS-HUMAN-DECISION: [count]

Triage table:
| # | Title | Current labels | Classification | Reason | Action |
|---|---|---|---|---|---|
| [n] | [title] | [labels] | [classification] | [1-line reason] | [proposed action] |

Proposed IMPLEMENT sequence:
1. #[n] — [title] — [why first]
2. #[n] — [title] — [dependency note]
...

Phase gate flags:
[any PHASE_GATE warnings]

Cross-issue dependencies:
[explicit dependency graph if any]

Doctrine violations detected:
[any CLOSE recommendations due to out-of-scope or stack violations]

NEEDS-HUMAN-DECISION items:
[list with precise question for owner]

Recommended next step:
- Approve sequence and proceed to /execute #[first issue]
- OR review NEEDS-HUMAN-DECISION items first

STOP. Waiting for instruction.

Handoff guidance

  • After triage report is reviewed and a sequence is approved → hand first issue to /execute
  • NEEDS-HUMAN-DECISION items → stop until owner responds
  • CLOSE / MERGE-DUPLICATE actions → execute only after explicit approval

Guardrails / hard rules

  • No code changes. Triage does not touch implementation files.
  • No silent issue closes. All CLOSE and MERGE-DUPLICATE actions require explicit approval before execution (unless --apply flag is set by the invoker).
  • No invented priorities. Sequence recommendations must be based on phase gates, dependencies, or explicit doctrine — not arbitrary judgment.
  • Deduplicate aggressively. Duplicate issues waste sprint capacity. Merge them.
  • Never preserve out-of-scope noise. Issues requesting features on the Phase 1 out-of-scope list should be closed, not deferred indefinitely.
  • Respect NEEDS-HUMAN-DECISION. Do not resolve architectural decisions autonomously.
  • Stop at ~80% context window pressure. Output STOPPED_CONTEXT_LIMIT.

Example invocations

/triage
/triage source=issue-list scope=phase:1
/triage source=audit-report
/triage source=both scope=backend
/triage source=issue-list scope=all --apply

Связанные навыки

Looking for an alternative to triage or another community skill for your workflow? Explore these related open-source skills.

Показать все

openclaw-release-maintainer

Logo of openclaw
openclaw

Your own personal AI assistant. Any OS. Any Platform. The lobster way. 🦞

widget-generator

Logo of f
f

Создание настраиваемых плагинов виджетов для системы ленты новостей prompts.chat

flags

Logo of vercel
vercel

Фреймворк React

138.4k
0
Браузер

pr-review

Logo of pytorch
pytorch

Tensors and Dynamic neural networks in Python with strong GPU acceleration

98.6k
0
Разработчик