prompt-maker-studio
AGENTS.md7 starsChanged 18 days agoApache-2.0View on GitHub
FastAPI backend (backend/) + Next.js frontend (frontend/), Docker/Compose behind Caddy. The backend calling a user's own LLM provider via backend/app/services/llm_client.py is a product feature, unrelated to this agent-operating layer. Go to the source; don't re-derive or duplicate it here. Stack conventions: backend/AGENTS.md, frontend/AGENTS.md. Step-by-step playbooks (new endpoint, DB migration) live in .claude/skills/ — check there before deriving a procedure. Keep instruction content in AGENTS.md; the CLAUDE.md files are symlinks. make help lists every target. CI enforces a 90% backend coverage…
- Reads credentials
Where files are headingNew AGENTS.md files have more rules about keeping secrets out of the code, and fewer about code style.
Of the AGENTS.md files we read, the 31,439 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.
In this repository10 other files it gives its agents.
- backend/AGENTS.mdAGENTS.md
- frontend/AGENTS.mdAGENTS.md
- add-api-endpointSkill
- code-reviewSkill
- coding-agent-briefSkill
- db-migrationSkill
- find-skillsSkill
- impeccableSkill
- product-ownerSkill
- readme-syncSkill
The file
We split this AGENTS.md into the rules it gives an agent
3 rules. About 0 are common, 1 is distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 29
Keep instruction content in AGENTS.md.
- Very specific: line 55
Keep commit messages short and concise.
- Very specific: line 63
Keep diffs minimal and coherent — don't opportunistically restructure code you happened to touch.
≈ “Common” is an estimate: at least 1 in 5 of the 40,246 AGENTS.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
prompt-maker-studio — agent operating notes
FastAPI backend (backend/) + Next.js frontend (frontend/), Docker/Compose behind Caddy.
The backend calling a user's own LLM provider via backend/app/services/llm_client.py is a
product feature, unrelated to this agent-operating layer.
Where things are already documented
Go to the source; don't re-derive or duplicate it here.
| Need | Where |
| Documentation map and current vs. historical records | docs/README.md |
| Architecture and tech stack | README.md |
| REST API routes and contracts | docs/api.md |
| Contributor conventions (branches, commits, PRs, coding standards) | CONTRIBUTING.md |
| Local dev setup | docs/development.md |
| Deployment (VM, Docker, Caddy, backups) | docs/deployment.md |
| Docker Compose details | docs/docker.md |
| CI pipeline | .github/CI_PIPELINE.md |
| Auth system | docs/authentication.md |
| Saved-prompts feature | docs/saved-prompts.md |
| What's already built | product/FEATURES.md — check before calling something new |
| Why past decisions were made | product/DECISIONS.md — check before re-litigating |
| Vision / roadmap / backlog | product/VISION.md, product/ROADMAP.md, product/BACKLOG.md |
Stack conventions: backend/AGENTS.md, frontend/AGENTS.md. Step-by-step playbooks (new
endpoint, DB migration) live in .claude/skills/ — check there before deriving a procedure.
Keep instruction content in AGENTS.md; the CLAUDE.md files are symlinks.
Definition of done
- backend change →
make lint-backend && make test - frontend change →
make lint-frontend && make test-frontend - touches both, or before considering any larger change complete →
make ci-local
make help lists every target. CI enforces a 90% backend coverage gate — new backend code
needs tests.
Repository etiquette
The ongoing broader refactoring started with UI cleanup and simplification. For this effort:
- Make all refactoring changes on
dev. - Develop each new feature on a separate
feat/short-descriptionbranch created fromdev. - Merge completed feature branches back into
dev.
This workflow takes precedence over the main branch base and rebase guidance in
CONTRIBUTING.md for this effort. Otherwise, branch from main using
feat|fix|ref|docs|chore/short-description.
Commits follow Conventional Commits (fix(auth): reject expired reset tokens). Full rules
in CONTRIBUTING.md.
Keep commit messages short and concise. Never add a Co-Authored-By trailer.
Guardrails
- YOU MUST NOT read, print, write, or edit
.env— it holdsSECRET_KEY,LLM_ENCRYPTION_KEY, SMTP creds, and the Grafana admin password. This applies to.envonly;.env.exampleis a checked-in template and is fine to edit. - Never deploy or publish without explicit user confirmation.
- Keep diffs minimal and coherent — don't opportunistically restructure code you happened to
touch. Other coding standards:
CONTRIBUTING.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.