amirfish1/watchtower/compact-to-queue
Claude Code and agent skill by amirfish1/watchtower
- community
Checking that the SKILL.md file is still in the repository.
Use when a session's context is filling up, before running /compact, or when the user says "remember this for later" — distill the remaining work into WatchTower queue tickets instead of relying on compaction or memory.
Install
npx vskill@latest install amirfish1/watchtower/compact-to-queueSKILL.md
Compact to queue
A compaction summary is lossy memory; a queue ticket is durable state. When a long session is about to compact — or is ending with work still open — file the remaining work as WatchTower tickets so any future session (human, agent, or drain worker) can pick it up cold.
REQUIRED BACKGROUND: the watchtower skill covers filing, claiming, and closing tickets. This skill is the discipline of what to extract and when.
When to fire
- Context is ~70%+ full, or the user asks "when should we compact?"
- The user says "don't forget X", "put this on the list", "file it".
- A session ends with known open items, limitations, or deferred work.
1. Pick the queue
- Prefer the repo's existing queue (
wt ls/wt statusto see what's in use). - For a workstream that spans sessions, a dedicated queue reads better
(e.g. KIMI-FIXES for the kimi-engine workstream). Ask if unclear.
- Check for dupes first:
wt ls -q <QUEUE> --status all --json.
2. Extract only what's NOT done
File: remaining work, known limitations, deferred fixes, ideas the user approved but aren't built. Never file completed work — the ticket is a promise of future effort, not a log.
One item per ticket. If a note needs "and also", that's two tickets.
3. File each item self-contained
The reader has zero context. Every note must stand alone: the why, any measured numbers, and file:line pointers to where the work lives.
Field limits (hard clips, so don't fight them): title 200 chars, note 4000, text 24000, comments 24000. Title = preview, note = brief, text = full body.
Honest priorities: p1 = next thing to do, p2 = should do, p3 = icebox. --readiness ready only when the work is fully specified; otherwise needs-shaping.
4. Verify and report
Show the user the refs you filed. Those refs — not the session, not the compact summary — are now the source of truth for what's next.
After filing
Compacting is safe: everything of value now lives in the queue. The session can end. The next session resumes from wt ls -q <QUEUE>, not from memory.
Security
- Security scanPassed, 90 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingwatchtower/skills/compact-to-queue/SKILL.md
Popularity
- 7-day trend score
- 0
Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/amirfish1/watchtower/compact-to-queue)