ahundt/autorun/ar
Claude Code and agent skill by ahundt/autorun
- community
Checking that the SKILL.md file is still in the repository.
autorun control commands, stated as prose after $ar: status; allow/justify/find file-creation policy; ok/no/blocks/clear command guards; global variants; stop/estop; task tracking and pause; cache-miss gate; planexport settings; reload; help lists everything.
Install
npx vskill@latest install ahundt/autorun/arSKILL.md
The same SKILL.md was published earlier as majiayu000/claude-skill-registry-data/ar.
autorun commands (ar)
<purpose>
autorun's daemon dispatches these from plain prompt text; no slash menu or catalog entry is required for them to work. Type ar:<command> [args] (ar <command> and /ar:<command> are also accepted). Type ar:help for the live list in this harness's spelling, or ar:help <command> for one command.
</purpose>
<commands>
File-creation policy (AutoFile)
ar:status(ar:st) — show the current policy and settings.ar:allow(ar:a) — allow all file creation.ar:justify(ar:j) — require a justification for each new file.ar:find(ar:f) — modify existing files only (strictest).
Command guards
ar:no <pattern>— block a command pattern in this session.ar:ok <pattern> [N|5m|perm]— allow a blocked pattern: N uses, a
duration such as 5m, 2h30m, or 2d, or perm for the rest of the session. Default is one use, then auto-revoke.
ar:blocks— show active session blocks and allows.ar:clear— clear all session blocks and allows.- Global variants persist across sessions:
ar:globalno <pattern>,
ar:globalok <pattern> [N|5m|perm], ar:globalstatus, ar:globalclear.
Run control
ar:go <task>(ar:run) — start autonomous execution with three-stage
verification.
ar:gp <task>(ar:proc) — procedural mode with the wait process.ar:stop(ar:x) — graceful stop after the current task.ar:estop(ar:sos) — emergency stop, cancels the current autorun.
Task tracking
ar:task/ar:tasks— task, pause, prompting, and recovery status.ar:task pause [N] [duration] [reason]— bare pause defaults to five
minutes; a reason-only pause holds until AI recovery; an explicit count and duration may be combined.
ar:task resume— resume task enforcement.ar:task ignore <id> [reason]— mark one task ignored.ar:task prompts on|off|<N>— configure task-staleness prompting.ar:task recovery on|off|min <N>— configure repeated-Stop stale-task
recovery.
Cache-miss / compaction gate (off by default)
ar:cache— show status.ar:cache on|off [5m|1h|perm]— toggle,
optionally for a window.
ar:cache set ratio|read|age|full <value>— configure threshold axes.
Tokens 50k or .5M, percents 85%, durations 5m, 2h30m, or 2d.
ar:cache ok [N|5m|perm]— override the gate;ar:cache no— cancel
outstanding overrides.
Plan export
ar:pe [on|off|globalon|globaloff]— status or toggles; a per-project
pin beats the global default.
ar:pe dir <path>— set the export directory.ar:pe pattern <template>— set the filename pattern.ar:pe rejected [on|off|dir <path>]— rejected-plan export.ar:pe reset— restore defaults and clear project pins.
Admin
ar:reload— force-reload integration rules from config files.ar:help— list every command with what it does.
</commands>
Security
- Security scanPassed, 95 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 repositoryCheckingplugins/autorun/.codex-plugin/skills/ar/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/ahundt/autorun/ar)