environment-preflight

Installation
SKILL.md

Environment Preflight

Before assuming you can build, test, lint, migrate, or run a project's toolchain, inspect the repository to determine what technologies it uses, then probe the active shell environment to confirm direct access to every required tool. Never guess, fabricate tool output, or silently skip a check when a CLI is unavailable.

1. Identify Project Stack & Required Tools

Scan the repository root (and workspace/monorepo packages, if present) for manifests, lockfiles, and configuration files to map the tech stack:

  • Runtimes & Version Pins: .nvmrc, .node-version, .python-version, .ruby-version, .tool-versions, mise.toml, rust-toolchain.toml, go.mod, engines in package.json, requires-python in pyproject.toml.
  • Package Managers & Lockfiles: package.json + package-lock.json / pnpm-lock.yaml / yarn.lock / bun.lockb, pyproject.toml / uv.lock / poetry.lock / requirements.txt, Cargo.toml, go.mod, pom.xml, build.gradle*, Gemfile, composer.json.
  • Frameworks, Build & Quality Tools: Scripts and configs (e.g., scripts in package.json, Makefile, justfile, Taskfile.yml, tsconfig.json, vite.config.*, next.config.*, pytest.ini, ruff.toml, .eslintrc*, biome.json).
  • Infrastructure & External CLIs: Dockerfile, docker-compose.yml, terraform/, k8s/, .github/workflows/*, database/migration configs (e.g., prisma/, drizzle.config.*, alembic.ini), or cloud/platform configs (vercel.json, firebase.json, fly.toml).

From these files, build a concise list of required tools for the current task (e.g., runtime, package manager, test runner, linter/typechecker, container or cloud CLI) and optional tools.

2. Probe the Shell & Development Environment

Check what the agent can directly execute in the current terminal/shell:

Installs
13
First Seen
9 days ago