LegalIP & Licensing

License Comply

Use when auditing open-source dependency licenses in Python projects, generating compliance reports, or checking license risk in your codebase.

LLegalQuants·Legal·Apache-2.0

Library skill — the default version is maintained in GitHub; edits you make live in your own clone.

Use this skillDownload .zip
How does this work?
  • ChatGPT opens a new chat with the skill loaded. If it's too long for a link, it's copied to your clipboard — just paste.
  • Claude works the same way. To install it permanently, download the .zip and upload it under Claude → Settings → Capabilities → Skills (Pro/Team/Enterprise).
  • Copy prompt copies the skill so you can paste it into any assistant, including Grok.

license-comply

When to Use

Audience and Work Shape

Audience: non-lawyer engineers running a CI-gated dependency scan, and the in-house counsel / open-source review function those engineers escalate to. Not intended as a stand-alone clearance tool for engineering teams without a counsel-in-the-loop on flagged findings.

Work shape: Pattern-Matched Review for license detection (license string -> SPDX -> policy band). The "AI executive summary" and "plain-English explanations" features are advisory framing, not judgment work, and must not be relied on as legal advice or as a substitute for counsel review.

Scope and Legal Use

This skill provides legal support, not legal advice. A "low risk" classification means "the detected SPDX identifier is on the configured allow-list" — it does not mean "this dependency is safe to ship in your product," "this license is compatible with the rest of your stack," or "your use of this code is compliant."

Three legal failure modes this skill explicitly addresses:

  1. Legal advice vs. legal support. The report is advisory. Risk bands, remediation text, plain-English explanations, and the optional AI executive summary are detection output, not legal conclusions. They do not constitute legal advice and must not be circulated to leadership as a clearance.
  2. Privilege implications. CI-generated reports are discoverable artefacts by default. If counsel wants the manual-review layer to sit inside privilege, counsel must mark up the report under privilege themselves; running the CLI in CI does not create work product. Treat raw CI artefacts as non-privileged.
  3. Accountability gap. The tool is CI-gated and aimed at non-lawyer engineers. Without an explicit gate, an engineer can merge against a license violation without legal review — the --fail-on=high flag exits non-zero but can be silently bypassed in CI configuration, and the AI executive summary lets leadership consume a license posture with no counsel in the loop. Review required by qualified open-source / IP counsel before merging changes flagged Medium or High. The CI gate is not a substitute for that review.

How It Works

Core Functionality

Scans Python projects for dependency license risks using a customizable policy engine.

Features

Installation

pip install license-comply

Usage

# Scan current directory
license-comply scan .

# CI mode (exit non-zero on violations)
license-comply scan . --fail-on=high

# Generate HTML report
license-comply scan . --format=html --output=report.html

# Custom policy
license-comply scan . --policy=/path/to/policy.yaml

Policy Engine

Define your organization's license policy:

allowed_licenses:
  - MIT
  - Apache-2.0
  - BSD-3-Clause

restricted_licenses:
  - AGPL-3.0
  - CC-BY-NC-4.0

flagged_licenses:
  - GPL-2.0: "Consult legal before use"

Confidence Bands

Output is advisory, not legal advice, regardless of band.

Out of Scope

Escalation

Stop, do not merge, and route to qualified open-source / IP counsel when:

When escalating, include: package name, version pin, license text as detected, dependency-graph location (direct vs. transitive, parent package), and the business use of the package.

Limitations

QA Remediation (LegalQuants, 2026-05)

Original author: Sam Clearwater. LegalQuants remediated this skill on 2026-05 against the Legal Skill Design Framework, addressing the three legal failure modes (legal advice vs. legal support, privilege implications, accountability gap), adding explicit audience / work shape / confidence-band / out-of-scope / escalation sections, and inserting an explicit "qualified open-source / IP counsel review required before merging Medium or High findings" gate to close the accountability gap for non-lawyer engineers running the CI tool. Technical content (installation, usage, policy engine) is unchanged. Output remains advisory, not legal advice.