ai-inventory
EU AI Act per-system inventory — track each AI system's role (provider, deployer, importer,…
Open source license compliance check for a dependency list, a single library, or outbound code. Use when reviewing a manifest, SBOM, or repo for copyleft obligations and license compatibility, when asked whether a library can ship, or when preparing code to be open-sourced.
$ npx -y skills add anthropics/claude-for-legal --skill oss-review --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/oss-reviewContext preview
The summary Claude sees to decide when to auto-load this skill.
Open source license compliance check for a dependency list, a single library, or outbound code. Use when reviewing a manifest, SBOM, or repo for copyleft obligations and license compatibility, when asked whether a library can ship, or when preparing code to be open-sourced.
name: oss-review description: > Open source license compliance check for a dependency list, a single library, or outbound code. Use when reviewing a manifest, SBOM, or repo for copyleft obligations and license compatibility, when asked whether a library can ship, or when preparing code to be open-sourced. argument-hint: "[file path to manifest / SBOM | package name | repo path | paste text]"
Runs an open source license compliance check against the practice profile in `~/.claude/plugins/config/claude-for-legal/ip-legal/CLAUDE.md`. Classifies dependencies by license family, maps obligations to the deployment model, flags license-unknown and non-OSI-posing-as-OSS packages, and recommends actions — comply, replace, remove, seek legal review, seek commercial license.
1. **Load `~/.claude/plugins/config/claude-for-legal/ip-legal/CLAUDE.md`.** If placeholders present, stop and prompt: "Run `/ip-legal:cold-start-interview` first — I need to learn your practice profile (and OSS policy, if any) before I can review." If the practice profile points at an uploaded OSS policy, read that too — it is the source of truth for accepted / review / banned licenses on this team.
2. **Establish the scope:** a dependency list (package.json, requirements.txt, go.mod, Gemfile, Cargo.toml, pom.xml, SBOM), a single library, or outbound code the team is preparing to open-source. If the user passed a path, infer from the file; otherwise ask.
3. **Establish the deployment model** before classifying obligations — SaaS, distributed binary, internal only, or embedded. The same dependency list triggers different obligations depending on this.
4. **Follow the workflow below.** In particular:
5. **Output the memo** per the template below — work-product header first, bottom line, top-of-memo flags, per-package blocks grouped by severity, jurisdiction note, outbound check (if applicable), approval routing.
6. **Respect the decision posture.** When a copyleft-trigger analysis turns on a contested question (AGPL's "interacts over a network," GPL-3.0's "conveying," LGPL linking scope), flag for attorney review and surface the factors cutting both ways. Anything flagged as strong copyleft or license-unknown goes to an attorney before the dependency ships or the code is released.
/ip-legal:oss-review ~/code/my-project/package.json /ip-legal:oss-review ~/code/my-project/requirements.txt /ip-legal:oss-review redis /ip-legal:oss-review ~/code/my-project # repo root — scan all manifests
---
OSS clearance requests usually come in via a ticketing system. Connected to Jira, Linear, or Asana, this skill can: monitor incoming OSS requests, respond with guidance directly in the ticket (flagging incomplete info, asking for the repo link, returning the license-family classification), and track clearance status across requests.
Without a connector, paste the ticket or describe the request and I'll handle it one at a time. See `CONNECTORS.md` at the repo root for how to add a ticketing connector.
**Matter context.** Check `## Matter workspaces` in the practice-level CLAUDE.md. If `Enabled` is `✗` (the default for in-house users), skip the rest of this paragraph — skills use practice-level context and the matter machinery is invisible. If enabled and there is no active matter, ask: "Which matter is this for? Run `/ip-legal:matter-workspace switch <slug>` or say `practice-level`." Load the active matter's `matter.md` for matter-specific context and overrides. Write outputs to the matter folder at `~/.claude/plugins/config/claude-for-legal/ip-legal/matters/<matter-slug>/`. Never read another matter's files unless `Cross-matter context` is `on`.
---
Tell the user what licenses are in their dependency tree, what obligations those licenses trigger given how the code will be deployed, and what to do about each one. The output is a memo the lawyer (or the engineer with attorney access) can act on — comply, replace, remove, seek legal review, seek commercial license.
**This is a first-pass classification.** Copyleft analysis depends on the deployment model, the degree of linking, the jurisdiction, and sometimes on legal questions that have not been tested in court (notably AGPL's "interacts over a network," GPL-3.0's patent clause). For anything that classifies as strong copyleft or license-unknown, an attorney evaluates before the dependency ships or the code is released. The skill reports what it found; the lawyer decides what to do.
**Before scanning dependencies, read `~/.claude/plugins/config/claude-for-legal/ip-legal/CLAUDE.md`.** If it is missing or still contains placeholders, stop and run `/ip-legal:cold-start-interview`. The practice profile tells you:
If the practice profile has an OSS policy uploaded, read that too — it is the source of truth for which licenses the team accepts, which trigger review, and which are banned.
Ask (or infer from what the user provided):
> What are we reviewing? > > 1. **A dependency list** — `package.json`, `requirements.txt`, `go.mod`, `Gemfile`, `Cargo.toml`, `pom.xml`, an SBOM (SPDX / CycloneD
Reference agents, skills, and data connectors for the legal workflows we see most — in-house commercial, privacy, product, corporate, employment, litigation, regulatory, AI governance, IP, and the learning side of the practice (law school clinics and
EU AI Act per-system inventory — track each AI system's role (provider, deployer, importer,…
Run an AI impact assessment — structured intake, risk analysis, regulatory classification per…
Run the cold-start interview — learns your AI governance practice and writes…
Guided customization of your AI governance practice profile — change one thing without…
Manage matter workspaces — new, list, switch, close, or detach (practice-level).…
Keep the AI policy current with practice — weekly sweep of saved AIAs, triage results, and…