validate-tools 0.1.0

Local CLI for fullstack code validation — import directions, migrations, secrets, and more

validate-tools

Local CLI for fullstack code validation. Runs the same checks as the MCP server without the token overhead — pipe files in, get JSON out.

Installation

pip install validate-tools
# or
uv tool install validate-tools

Usage

validate-tools [--human] [--pretty] COMMAND [OPTIONS]

When stdout is not a TTY (piped to a script or agent), JSON is emitted automatically. No --json flag needed.

Global flags:

Flag	Description
--human / -H	Force rich/table output even when piped
--pretty	Indent JSON output (default: compact)

Exit codes: 0 = clean · 1 = warnings (with --strict) · 2 = violations

---

Commands

imports — Clean Architecture import directions

grep -rn "^from\|^import" src/ --include="*.py" | validate-tools imports

Rules: domain → no project imports; application → domain only; infrastructure → domain + application; presentation → all layers.

---

commits — Conventional Commits format

git log --format="%H %s" origin/main..HEAD | validate-tools commits

Checks type(scope)?: description format (required) and ≤72-char subject (recommended). Allowed types: feat fix docs chore refactor test ci perf build style revert

---

migration — Alembic migration safety

validate-tools migration alembic/versions/*.py

Flags: drop_column, drop_table, rename_table, rename_column, add_column with nullable=False and no server_default (required); alter_column, execute (recommended).

---

env — Settings ↔ .env.example completeness

validate-tools env --settings src/config/settings.py --example .env.example

Every UPPER_SNAKE_CASE field in class *Settings* must have a matching KEY= entry in .env.example.

---

tests — pytest test name quality

validate-tools tests test/unit/test_auth.py

Flags duplicate test_* names (required) and names with fewer than 3 tokens after test_ (recommended).

---

logs — Logging anti-patterns

validate-tools logs src/application/use_cases/auth_use_case.py

Flags print() calls (required) and f-strings inside logger.*() (recommended).

---

coverage — Per-layer coverage thresholds

pytest --cov=src --cov-report=xml
validate-tools coverage coverage.xml

Thresholds: domain ≥90% · application ≥85% · infrastructure ≥65% · presentation ≥55%

---

supply-chain — Dependency manifest risks

validate-tools supply-chain pyproject.toml   # or package.json

Flags VCS/URL/local-path sources and wildcard versions (required); pre-release versions (recommended).

---

sensitive-logging — Sensitive data in logs

validate-tools sensitive-logging src/application/use_cases/auth_use_case.py

Flags passwords, tokens, secrets, API keys, and card numbers passed to log.*() or print().

---

secrets — Hardcoded credentials

validate-tools secrets src/config/settings.py

Detects Stripe, Slack, GitHub, Google, and AWS key literals; JWT tokens; and sensitive variable assignments (password = "...", secret = "...").

---

Batch mode (run) — for AI agents

Run multiple validators in a single invocation. Accepts a JSON config on stdin and returns a JSON array — one report per check.

cat <<'EOF' | validate-tools run
{
  "imports":  "<output of grep -rn ...>",
  "commits":  "<output of git log --format='%H %s' ...>",
  "supply_chain": "<content of pyproject.toml>",
  "secrets": [{ "filename": "settings.py", "source": "<file content>" }],
  "logs":    [{ "filename": "auth.py",     "source": "<file content>" }]
}
EOF

Config keys (all optional):

Key	Value
imports	grep output (string)
commits	git log output (string)
migration	migration file content (string)
coverage	coverage.xml content (string)
supply_chain	pyproject.toml or package.json content (string)
env	{"settings_source": "...", "env_example": "..."}
tests	[{"filename": "test_foo.py", "source": "..."}]
logs	[{"filename": "foo.py", "source": "..."}]
sensitive_logging	[{"filename": "foo.py", "source": "..."}]
secrets	[{"filename": "foo.py", "source": "..."}]

For per-file checks, a plain string is also accepted (filename defaults to source.py).

---

Output format

Every command emits the same JSON schema:

{
  "analysis": "validate_import_directions",
  "status": "clean | warnings | violations",
  "total_items": 42,
  "required_count": 0,
  "recommended_count": 0,
  "summary": "All 42 file(s) respect the layer dependency rules.",
  "findings": [
    {
      "rule_id": "backend/imports/domain-no-infrastructure",
      "severity": "required | recommended",
      "location": "src/domain/entities/user.py:12",
      "message": "Domain layer must not import from infrastructure layer",
      "hint": "Remove the cross-layer import. Domain may only depend on: domain only."
    }
  ]
}

Errors also follow a consistent schema when piped:

{"status": "error", "error": "git_log is empty — run: ...", "analysis": "validate_commit_messages"}

License

MIT