Skip to content

AGENTS.md

Latest commit

 

History

History
49 lines (33 loc) · 2.46 KB

AGENTS.md

File metadata and controls

49 lines (33 loc) · 2.46 KB

AGENTS.md

Project Overview

  • Monorepo with independent tools under packages.

Setup Commands

WIP

Build and Smoke Test Commands

WIP

Code Style and Change Scope

  • Keep changes scoped to one package unless a cross-cutting workflow change is required.
  • Prefer minimal edits; avoid broad reformatting of config or generated files.
  • When adding a package, follow existing layout:
    • packages//README.md

Architecture and Key Files

WIP

CI and Release Conventions

  • Multi-architecture image publishing is required:
    • linux/amd64
    • linux/arm64
  • If release to Docker Hub, README sync source of truth is required.

TypeScript Scripts

  • Standalone scripts: every operational script in scripts/ must be directly runnable as its own entrypoint and must not depend on application/package implementation files. The only local-code exception is importing helper or utility modules from scripts/lib/.
  • Shebang: start every .ts or .mts file that is entrypoint with #!/usr/bin/env bun.
  • Runtime: Bun — use node:* built-ins, never install third-party CLI libs (e.g., commander, yargs). Use node:util parseArgs for argument parsing.
  • Imports: use node: prefix for all Node built-ins (e.g., node:fs, node:path)
  • Shared code: put reusable helpers and utilities in scripts/lib/ — operational scripts import from there. Organize lib/ files by domain (e.g., dockerhub.ts for Docker Hub API interactions, utils.ts for general utilities).
  • Structure: break scripts into small, focused functions — one function per logical step. The main flow should read like a sequence of named steps, not a single monolithic block.
  • Logging: use node:util styleText for colored console output (e.g., styleText('green', '✓ Done'), styleText('red', 'Error: ...')). Use the logStep, logSuccess, logError, logWarn, logInfo helpers from lib/helpers.ts when available. Never use raw ANSI escape codes or chalk/picocolors.
  • Comments: every script must have a top-level doc comment explaining what it does, common usage, and any required env vars. All utility functions should have a doc comment describing their purpose, inputs, and outputs.
  • No secrets in code: reference env vars or .env files — never hardcode keys, tokens, or passwords
  • Exit codes: use non-zero exit codes for failures; print a clear error message before exiting

References

  • Root overview: README.md
  • Packages: packages/*/README.md