setup
Skill0 starsChanged 38 days agoMITView on GitHub
Install this app's workflow pieces. Write what is missing. Rewrite what no longer matches the app.
191 other projects use a version of this.
Adopted by
Most recently taken up by
- GoogleCloudPlatform/gemini-cloud-assist-mcptheir own version, first seen Nov 15, 2026
- finolabs/vibecodetheir own version, first seen Nov 15, 2026
- quickchatai/quickreviewtheir own version, first seen Nov 15, 2026
- tomw1808/flash-agentstheir own version, first seen Nov 15, 2026
- felipeam86/pipe-skillstheir own version, first seen Nov 15, 2026
- oss-invisible/playwright_mcptheir own version, first seen Nov 15, 2026
Used with
Projects with it also use add-voice-transcription, add-discord and 10 more.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
juancruzmunozalbeloSkilladd-voice-transcription
Add voice message transcription to NanoClaw using OpenAI's Whisper API. Automatically transcribes WhatsApp voice notes so the agent can read and respond to them.
Also in 5 projects that use this
MicrosoftSkilladd-discord
Add Discord bot channel integration via Chat SDK.
Also in 11 projects that use this
MicrosoftSkilladd-slack
Add Slack channel integration via Chat SDK.
Also in 11 projects that use this
juancruzmunozalbeloSkilladd-telegram-swarm
Add Agent Swarm (Teams) support to Telegram. Each subagent gets its own bot identity in the group. Requires Telegram channel to be set up first (use /add-telegram). Triggers on "agent swarm", "agent teams telegram", "telegram swarm", "bot pool".
Also in 7 projects that use this
- +8 more from microsoft
Other versions191 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 191
- anthropics/knowledge-work-plugins★ 36ksame name
- garrytan/gbrain★ 32ksame name
- nanocoai/nanoclaw★ 31ksame name
- gavrielc/nanoclaw★ 31ksame name
- qwibitai/nanoclaw★ 31ksame name
- Q00/ouroboros★ 6.4ksame name
- pytorch/executorch★ 5.2ksame name
- aws/agent-toolkit-for-aws★ 3.5ksame name
- agenticnotetaking/arscontexta★ 3.5ksame name
- spotify/portal-ai-plugins★ 2.8ksame name
181 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
The file
We split this skill into the rules it gives an agent
8 rules. 0 are common, 2 are distinctive, and 6 are very specific to this project.
0
Common
Most of the 170 edited versions we read keep these.
2
Distinctive
This author's own choices. Worth a look.
6
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 15
Do not copy hard, fast, debug, think, or clean into the repo.
- Very specific: line 17
No app yet (no manifest and no start script): reply ⚠️ no app and stop.
- Very specific: line 25
Write .cursor/skills/radar/SKILL.md with disable-model-invocation: true.
- Very specific: line 31
Run the commands in this skill.
- Distinctive: line 34
No diff: that service already matches main.
- Very specific: line 38
All of that is green: reply ✅ and stop.
- Very specific: line 41
If this repo has a GitHub CI on main, add the command that prints whether that run for this commit is green, and add this rule: The GitHub …
- Distinctive: line 57
Use the same proof as Radar.
“Common” is measured: kept by at least 6 in 10 of the 170 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Setup
Read the repo.
For each piece below, look in the repo. Missing, write it. Already there, check that it is perfectly up to date and matches this section exactly. If it does not, rewrite it.
Which services exist: look in this order. Files in the repo. The GitHub page of this repo, its site link and its deploy checks. Host folders on disk that git does not commit. A name in .gitignore is not proof. The folder is. The first proof wins. All three empty: the host is unknown.
Do not copy hard, fast, debug, think, or clean into the repo.
No app yet (no manifest and no start script): reply ⚠️ no app and stop. Do not invent one.
Reply ✅ when nothing changed, or ⚙️ plus the piece names you changed. Use ⚠️ for a problem or something the user must do.
Radar
Learn every service this app deploys. Use the proof above. A service with no proof is out. Write one ready command per check. Take it from the repo when the repo has one. When the host is known and the repo has none, look up that host's command in its docs. Do not run those commands. Do not install a tool. Each command must exit on its own.
Write .cursor/skills/radar/SKILL.md with disable-model-invocation: true. Description: After main moves, check that everything this app deploys matches main.
Under Commands, one block per service: the paths that belong to it, the command that prints the live commit, the command that prints the deploy state, the command that prints the logs since that deploy, and, when it has a public address, the command that requests it and prints the status. That command exits non-zero when the address cannot be reached or the status is 5xx. Use the health address the repo already exposes when it has one. These commands are specific to this repo. The rules below are copied word for word.
Rules:
- Run the commands in this skill. You own the current
origin/main. Fetch at the start and before the verdict. If main moved, reply🛑and stop. That emoji is the whole reply. - A check cannot run because a token or key is missing: reply
⚠️and name that key. The user must add it or the check cannot run. Then stop. Do not wait. - For each service, diff the live commit with main, only on that service's paths.
- No diff: that service already matches main. Skip it.
- A diff: wait until that deploy is live, or it has failed. If main moves while you wait, reply
🛑and stop. - The wait ends and that commit never deployed: the deploy filter is wrong. Fix it.
- For each service that just deployed: run its log command and its public address command when the skill gives one.
- All of that is green: reply
✅and stop. That emoji is the whole reply. Do not invent✅. - Something failed and the cause is clear: fix it. The cause is unknown: use the debug skill, then fix it.
If this repo has a GitHub CI on main, add the command that prints whether that run for this commit is green, and add this rule: The GitHub CI run on main for this commit is green. If it has none, add neither.
Then one automation named Radar - <app name>. Trigger: a pull request is merged into main, so main updates. The prompt is exactly Follow the radar skill at .cursor/skills/radar/SKILL.md. A newer merge cancels the running one. Create it yourself when you can. When you cannot, reply ⚠️ Add this automation if it is missing. Details: then three bullets: the name, the trigger, and the prompt.
Deploy
Same services as Radar. Same paths. Do not write a skill. Change the repo so a service ships when main updates and its paths changed. A commit that touches none ships nothing.
The host already builds on a push: put the skip in its config file. Missing, create it. Already there, fill what is missing. Exit 0 skips. Any other exit builds. No parent commit builds. Look up that host's switch in its docs. Do not add a workflow that deploys it again.
The service ships only when a command runs: one workflow per service, on push to main, limited to its paths. The command is the one the repo already runs. Missing, create it. Already there, fill what is missing.
Paths are the service root, the workspace packages it uses, the lockfile, and the file you write. A shared path ships both. Docs and skills ship nothing. When unsure, ship.
A build command that also deploys another service: split them. Each keeps its own filter. Leave preview deploys as they are.
Use the same proof as Radar. The first proof names the host. Write that host's filter. Missing, create it. Do not create a cloud project.
Files, the GitHub page, and the host folders are all empty: do not pick a host. Reply ⚠️ and ask where production deploys.
You cannot prove the host is connected, a secret is missing or you could not check it, or a dashboard also deploys this repo: write the file anyway. Reply ⚠️ and bullets: the one thing the user must do, the secret name, and where it goes. The skip lives only in a dashboard: name that switch in those bullets. Do not ask how.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

