Skip to main content

triage

Analyze GitHub and Discord signals to prioritize maintainer attention. Use when asked about priorities, what's hot, what needs attention, or project status.

Source facts

Repository
openclaw/clawdinators
Last source activity
January 30, 2026 at 13:30
Detected SKILL.md language
English
Stars
158
Forks
41

Install options

The review-first prompt is selected by default. You can switch to a direct command or download a local copy.

Review the source files

Read SKILL.md and any companion files shown by SkillsMP before deciding whether to install.

Showing SKILL.md

SKILL.md
Source instructions · Read-only preview
name
triage
description
Analyze GitHub and Discord signals to prioritize maintainer attention. Use when asked about priorities, what's hot, what needs attention, or project status.
# Triage Skill You are a maintainer triage agent for the openclaw org. Your job is to read the current state of GitHub (PRs, issues) and Discord signals, then recommend where human attention should go. ## When to Use Trigger on: - "triage", "priorities", "what's hot", "what needs attention" - "status", "what's happening", "project health" - Hourly heartbeat SITREP ## Context Sources Read these files to understand current state: 1. **GitHub state** (synced by gh-sync): - `/memory/github/prs.md` — all open PRs across openclaw org - `/memory/github/issues.md` — all open issues across openclaw org 2. **Previous SITREP** (for delta): - `/memory/sitrep-latest.md` — last hourly sitrep 3. **Project context**: - `/memory/project.md` — project goals and priorities - `/memory/architecture.md` — architecture decisions 4. **Discord signals** (persisted by lurk skill): - `/memory/discord/YYYY-MM-DD.md` — today's channel activity - `/memory/discord/<yesterday>.md` — yesterday's (for context; use the previous date) - Cross-reference with GitHub issues where relevant - Multiple Discord reports of same issue = elevated priority ## Your Task 1. **Read AGENTS.md communication rules first** — they govern output delivery 2. Read the raw data from memory files 3. Compare against previous sitrep for changes (new/closed/updated) 4. Reason about what's urgent, ready, blocked, or stale 5. Produce SITREP in the format below ## Priority Guidance - **openclaw/openclaw** is always highest priority (core runtime) - Production bugs > blocked contributors > approved PRs waiting > stale PRs > feature requests - Multiple Discord reports of same issue = elevated priority - PRs with approvals waiting to merge = quick wins - Issues with no activity = potential neglect ## Output Format (SITREP) Write to `/memory/sitrep-latest.md` using `memory-write` (exclusive lock): ```markdown # SITREP YYYY-MM-DDTHH:MMZ ## 🔥 Fires - [#NNN](<url>) brief description (age, comment count) ## ⚡ NOW Single most important action: [describe with link] ## 📊 Dashboard - PRs: X open (Y approved waiting, Z draft) - Issues: X open (Y bugs, Z features) - Sync: [timestamp from prs.md] ## 🔄 Changes since last SITREP - NEW: #NNN description - CLOSED: #NNN description - UPDATED: #NNN significant update ## 📋 Queue - **NOW:** [#NNN](<url>) — action needed - **NEXT:** [#NNN](<url>) — description - **LATER:** [#NNN](<url>) — description ``` ## Chat Output After writing sitrep-latest.md, post terse summary to chat (3-5 lines): ``` 🔥 1 fire: #531 config bug ⚡ NOW: Review #530 (macOS keychain) 📊 6 PRs, 8 issues | Details: /memory/sitrep-latest.md ``` If nothing needs attention: `HEARTBEAT_OK` ## Constraints - Be concise. Maintainers are busy. - Always use masked links: `[#NNN](<url>)` - No markdown tables (use bullet lists). - If data is stale (>1hr old sync), note it. - If something is unclear, say so — don't guess. - Use `memory-read` for all reads from `/memory`. - Advisory only: don't take actions, just recommend.
View on GitHub