Full-Stack AI Agent Template is an interactive CLI tool that generates production-ready FastAPI + Next.js projects with AI agents, RAG, and 20+ enterprise integrations. Uses Cookiecutter templates with Jinja2 conditionals.
Example: "Build meaningful technology while maintaining balance across health, relationships, and personal growth." See @Goals/2. Monthly Goals.md for this month's priorities. Priority: #priority/high, #priority/medium, #priority/low Status: #active, #waiting, #completed, #archived Context: #work, #personal, #health, #learning, #family Skills are invoked with /skill-name or automatically by Claude when relevant. Skills and agents use session task tools to show progress during multi-step operations: Session tasks are temporary progress indicators—your actual to-do items remain as markdown checkboxes in daily notes. Productivity Coach (/output-style coach)…
- Build: ./bun run build - Compiles TypeScript to JavaScript - Watch mode: ./bun run watch http - Watches for changes and rebuilds automatically - Run server: ./bun run start stdio - Starts the MCP server using stdio transport - Checks: ./check - All-in-one command to lint and test. Run before every commit. Covers TypeScript (biome, tsc, tests) and Python (ruff format + ruff check in ingest/). Python checks are skipped with a warning if uv is missing. - Release…
Guidance for Claude Code sessions working in this repo. This is supabase/postgres: Supabase's own build/packaging repo for its PostgreSQL distribution. It does not contain a fork of PostgreSQL's C source.
This file is the short Claude Code entrypoint for this repository. Always use the Babysitter skill to fulfill user requests in this repo, but not agent-to-agent delegated requests. The detailed rule set lives in docs/agent-reference/process-authoring.md.
This repo IS the USB drive content, minus large dependencies. The relationship: The repo is NOT a "build tool" or "generator" — it IS the USB structure. setup.sh only fills in large deps that can't go in git. After setup.sh, the portable/ folder is directly copyable to a USB drive. Distribution forms: 1. Portable USB (portable/): Run from USB on existing Mac/Windows, zero install. This is the only form CI publishes (Windows full zip; Mac runs setup.sh on first launch).…
The root CLAUDE.md imports the repository constitution from AGENTS.md; do not reread it when it is already present in active context. Background subagents cannot prompt for missing tool permissions. If a tool call is denied because it is not allowed in .claude/settings.json, fail fast and report the exact tool, command, and missing permission. Do not retry permission-denied calls in a loop. Run commands in the foreground when interactive approval may be required. Use the current provider's native subagent facility for…
Follow the canonical TypeScript compiler-ownership rules in AGENTS.md. Use repository/package typecheck and build scripts; yarn tsc ... is safe from the repository root and every TypeScript-owning workspace, and delegates to the native compiler runner. For a direct ad hoc check, use node scripts/workspaces/runTypeScriptCli.mjs .... Do not run bare tsc, npx tsc, node_modules/.bin/tsc, or typescript/bin/tsc: compilation must use the centrally resolved native TypeScript 7 compiler, while the typescript package remains TypeScript 5.9 only for programmatic API compatibility.
Agent-facing notes for local development. last30days-cn researches recent Chinese-platform discussion across Weibo, Xiaohongshu, Bilibili, Zhihu, Douyin, WeChat public accounts, Baidu, and Toutiao. The default window is the last 30 days.
A Claude Code skill (aso-appstore-screenshots) that guides users through creating high-converting App Store screenshots. It is invoked via the /aso-appstore-screenshots slash command from within a user's app project. Four files + one asset make up the skill:
claude-reflect is a Claude Code plugin that implements a two-stage self-learning system: 1. Capture Stage (automatic): Hooks detect correction patterns in user prompts and queue them 2. Process Stage (manual): /reflect command processes queued learnings with human review and writes to CLAUDE.md files The plugin registers via .claude-plugin/plugin.json: - Hooks are defined in hooks/hooks.json - Commands (skills) are markdown files in commands/ - SKILL.md provides context when the plugin is active
Run the binary: ./target/release/agent-desktop snapshot --app Finder -i The E2E harness drives the release binary against a real SwiftUI/AppKit fixture and asserts every effect by independent observation (never the command's own ok:true), covering every ref action in both headless and --headed mode. See tests/e2e/README.md. The repo ships a pre-commit hook at .githooks/pre-commit that runs cargo fmt --check, cargo clippy --all-targets -- -D warnings, and cargo test --lib --workspace against staged Rust changes. Wire it up once after cloning: Bypass for…
Single source of truth for what is structurally legal in the wiki. YAML in schema/ and policy/ is the contract; loader.py is the Python access API consumed by tools/lint.py and tools/research_wiki.py. No codegen step — edit YAML, restart the tool. Every change below is YAML-only, zero Python change. Edit, save, done. - New entity type — add a top-level block (<kind>: with dir, fields, optional lifecycle, optional terminal: true) and create templates/<kind>.md.tmpl with the body section skeleton. - New field…
All project guidelines live in AGENTS.md. Read that file — it is the single source of truth. This file exists only so Claude Code automatically picks up the project context. Do not duplicate content here.
A file Claude Code reads at the start of every session. It holds the commands, conventions and warnings the agent needs for this project.
Where does it go?
At the repository root. Claude Code also reads CLAUDE.md files in subdirectories when it works there.
What should it contain?
Build and test commands, the project's layout, conventions that aren't obvious from the code, and mistakes to avoid. Short files tend to work better than long ones.
CLAUDE.md or AGENTS.md?
Claude Code reads CLAUDE.md; most other agents read AGENTS.md. Many projects keep one and point the other at it.
yoanbernabeuCLAUDE.md
