llm-settings-template
CLAUDE.md0 starsChanged 39 days agoMITView on GitHub
This repo stores shared LLM context, task docs, and product docs for [project-name]. Code repos (app/, landing/, nginx/) are separate git repos and are ignored by the workspace repo. This workspace is local-only and is not deployed. Docker-only for code work: tests, lint, dependency operations, and service commands run through Docker. Git runs locally. After Docker builds, clean only dangling layers when needed: Never run docker system prune -a, docker volume prune, or remove named volumes unless explicitly requested. Always…
- Pipes a download into a shell
- Reads credentials
- Deletes or force-pushes
- Commits and pushes
In this repository5 other files it gives its agents.
- AGENTS.mdAGENTS.md
- complete-taskSkill
- create-taskSkill
- review-taskSkill
- start-taskSkill
The file
We split this CLAUDE.md into the rules it gives an agent
7 rules. About 0 are common, 2 are distinctive, and 5 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
5
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- could be automatic: a setting could enforce it
- Distinctive: line 26
After Docker builds, clean only dangling layers when needed:
- Very specific: line 32
Never run docker system prune -a, docker volume prune, or remove named volumes unless explicitly requested.
- Distinctive: line 34
Always search code with an explicit repo path:
- Very specific: line 44
Do not read .env, .env.*, or private key material (.ssh/, id_rsa*, *.pem, *.key).
- Very specific: line 46
Never run rm -rf, git push --force, git reset --hard, chmod 777, sudo rm, or curl/wget ...
- Very specific: line 50
Never hardcode or log secrets.
- Very specific: line 51
Before deploy: run pre-deploy checks from docs/runbook-prod.md and review the outgoing diff for secrets and debug leftovers.
≈ “Common” is an estimate: at least 1 in 5 of the 26,980 CLAUDE.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
[project-name] Workspace
This repo stores shared LLM context, task docs, and product docs for [project-name].
Code repos (app/, landing/, nginx/) are separate git repos and are ignored by the workspace repo. This workspace is local-only and is not deployed.
Source Of Truth
| Topic | File |
| Workspace map and repo commands | PROJECT_MAP.md |
| Task and docs rules | docs/folder-rules.md |
| Verification contract | docs/product/architecture/verification-contract.md |
/create-task, /start-task, /review-task, /complete-task flow | docs/workflows/task-lifecycle.md |
| Production deploy / rollback | docs/runbook-prod.md |
| Local development | docs/runbook-local.md |
| App context | app/CLAUDE.md, app/AGENTS.md |
| Landing context | landing/CLAUDE.md, landing/AGENTS.md |
| Nginx context | nginx/CLAUDE.md, nginx/AGENTS.md |
| Completed tasks | docs/done/short/ |
Rules
Docker-only for code work: tests, lint, dependency operations, and service commands run through Docker. Git runs locally.
After Docker builds, clean only dangling layers when needed:
docker image prune -f --filter "dangling=true"Never run docker system prune -a, docker volume prune, or remove named volumes unless explicitly requested.
Always search code with an explicit repo path:
Grep("pattern", path: "app")
Grep("pattern", path: "landing")
Grep("pattern", path: "nginx")Local Permissions
- Do not read
.env,.env.*, or private key material (.ssh/,id_rsa*,*.pem,*.key). .env.examplemay be read and edited.- Never run
rm -rf,git push --force,git reset --hard,chmod 777,sudo rm, orcurl/wget ... | bash.
Security
- Never hardcode or log secrets; secrets live only in untracked
.env, new variables go to.env.exampleas placeholders. - Before deploy: run pre-deploy checks from
docs/runbook-prod.mdand review the outgoing diff for secrets and debug leftovers. - On the server: inspect before modifying; do not touch firewall, sshd, TLS, or server
.envunless the task explicitly asks. - Full policy:
AGENTS.md→ Security.
SSH
ssh [your-server-alias]. Deploy and rollback: docs/runbook-prod.md. Local setup: docs/runbook-local.md.
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.

