validating-project

name	validating-project
description	Auto-detect and run formatters, linters, type checkers, and tests for the current project. Use when validating a project, running all checks, checking code quality before committing, or verifying the build passes.
context	fork
model	claude-haiku-4-5

Validating Project

Auto-detect project tooling and run validation steps in the correct order.

Process

1. Check ./CLAUDE.md for validation tools and project permissions
2. Auto-detect project type if needed (package.json, pyproject.toml, Cargo.toml, Makefile, go.mod)
3. Run validation pipeline: Format -> Lint -> Type Check -> Test
4. Stop immediately on failures
5. Report results

Tool Commands by Project Type

Make-based (check first -- overrides everything)

make format && make lint && make typecheck && make test

Or make validate if available.

Python

uv run ruff format .
uv run ruff check .
uv run ty check        # or uv run mypy, uv run pyright
uv run pytest

Node.js / TypeScript

pnpm run format        # or npx prettier --write .
pnpm run lint          # or npx eslint .
pnpm run typecheck     # or npx tsc --noEmit
pnpm test

Rust

cargo fmt
cargo clippy
cargo test

Go

gofmt -w .
go vet ./...
go test ./...

Auto-Discovery Logic

# Priority order
if [ -f "Makefile" ]; then
  # Check for make validate/format/lint/test targets
elif [ -f "pyproject.toml" ]; then
  # Python -- check for uv, poetry
elif [ -f "package.json" ]; then
  # Node.js -- check for pnpm, npm, yarn
elif [ -f "Cargo.toml" ]; then
  # Rust
elif [ -f "go.mod" ]; then
  # Go
fi

Error Handling

• Tool not found: Provide installation instructions
• Configuration missing: Suggest minimal setup
• Validation failures: Show specific fix recommendations
• Dependency conflicts: Suggest resolution strategies

CLAUDE.md Integration

Cache discovered validation info in the project's CLAUDE.md:

## Project Validation Tools

- **Format**: make format
- **Lint**: make lint
- **Type Check**: make typecheck
- **Test**: make test

Reporting Format

Read ~/.claude/references/status-marks.md and follow its conventions. Report one line per check, then a single plain-text summary line.

Example output:

❨`✓`❩ Trim trailing whitespace
❨`✓`❩ Fix end of file markers
❨`~`❩ Mixed line endings (skipped — none found)
❨`✗`❩ Markdownlint (3 errors)

After the list, add a plain-text summary with no marks — e.g. "all checks passed" or "1 failure: markdownlint". Skip reasons go in parens on the same line. Never collapse the per-item list into a single emoji'd summary.