What does this skill do, and when should you use it?
This is one of 14 skills bundled in the AI Legal Assistant repository for Claude Code, focused specifically on translating contracts into accessible language. It goes section by section, pairing each clause's original text with a plain-English explanation at roughly an 8th-grade reading level, and applies five flag types (e.g., WATCH OUT, DELIBERATELY CONFUSING) to language designed to obscure or that hides obligations. For the 5-10 most consequential clauses it adds expanded 'What This Really Means For You' callouts with concrete examples. The output is a structured Markdown report with a quick-reference summary, a glossary, and a flag summary. It explicitly does not constitute legal advice and should not replace an attorney.
Triggered by /legal plain <file>, it accepts a file path (read via the Read tool), a URL (fetched via WebFetch), or pasted contract text. It identifies the contract type, parties, and purpose, derives a short name, then produces for each clause: the exact original text, a 1-3 sentence plain-English translation, and applicable flags (DELIBERATELY CONFUSING, WATCH OUT, SURPRISINGLY BROAD, HIDDEN OBLIGATION, CONTRADICTS EXPECTATIONS). It prepends a one-page quick-reference summary (obligations, duration, exit routes, top risks) and writes a PLAIN-ENGLISH-[contract-name].md file containing the section-by-section translation, a glossary of the legal terms actually used, and a flag summary table with counts.
- A freelancer about to sign a client contract wants to understand what IP assignment and payment clauses actually commit them to
- A small business owner reviewing a vendor agreement without a lawyer's budget needs to identify the dangerous clauses first
- An individual signing a lease, employment contract, or SaaS agreement wants to know what they're really agreeing to
- A paralegal preparing a client-facing explanation of a contract in accessible language
- A founder comparing several contracts who wants readable summaries that surface hidden obligations
How do you install this skill?
- This is a static source review; no code was executed and actual translation/flag quality is unverified.
- Skill output is not legal advice; consult a licensed attorney for significant contracts.
- English-only output; suitability for Chinese contracts or non-English jurisdictions is undeclared, and mainland-China users may face reachability issues due to dependence on Anthropic services.
- README market statistics and '60-second review' marketing claims are unsourced and should not be treated as capability evidence.
- Repository license metadata is unknown; confirm licensing terms before commercial use.
- No confirmation before writing output files; be aware new Markdown files are generated in your working directory.
- Network access
- Local filesystem
Claude Code built-in Read and WebFetch tools
The skill ships as part of the AI Legal Assistant collection. The documented install is the one-line installer: curl -fsSL https://raw.githubusercontent.com/zubair-trabzada/ai-legal-claude/main/install.sh | bash — this installs all 14 skills, 5 agents, and the PDF scripts. Uninstall via uninstall.sh. The repo does not document a manual, standalone install of just skills/legal-plain; if you want only this skill, copying the skills/legal-plain/ folder into your Claude Code skills directory is the implied route, but the exact path is not officially documented.
tmp="$(mktemp -d)"
git clone --depth 1 https://github.com/zubair-trabzada/ai-legal-claude.git "$tmp"
mkdir -p ~/.claude/skills
cp -R "$tmp/skills/legal-plain" ~/.claude/skills/
rm -rf "$tmp"Generated from the source repository and skill path; it copies only this skill's folder. If the author's install steps above differ, follow those first. To scope it to one project, replace ~/.claude/skills with that project's .claude/skills.
How do you use this skill?
In Claude Code, run /legal plain <path-or-URL> or paste contract text inline, e.g. /legal plain ~/contracts/vendor-agreement.pdf. The skill reads the contract, translates and flags each clause, and writes PLAIN-ENGLISH-[contract-name].md next to the input file. You can also ask follow-up questions about specific clauses. Note this is a slash-command form; on other clients the trigger would need to be converted to a plain prompt.
What are this skill's strengths and limitations?
- Translates every section against the original text, including boilerplate where important terms often hide
- Five flag types plus expanded real-world callouts directly target misleading drafting and hidden obligations
- Produces a structured Markdown deliverable with summary, glossary, and flag counts, ready to share
- Includes an explicit legal disclaimer and instructs the model to explain meaning without giving advice
- Provides no legal advice and cannot replace an attorney for consequential contracts
- License is unknown, so commercial-use terms are unclear
- No test suite or evaluation data; translation and flag quality depends on the model with no third-party validation
- Built around Claude Code's slash-command syntax and Read/WebFetch tools; porting to other platforms requires rework
- Covers only one of the collection's 14 skills (plain translation) — it does not offer the risk scoring, NDA generation, or PDF reporting of sibling skills
How does this skill compare with similar options?
Side by side with related skills; every score comes from the same FSRS standard.
| Skill | FS score | Stars | Last updated | License |
|---|---|---|---|---|
| Plain English Contract Translator this page | 44 · Not recommended | ★ 1.8k | 7mo ago | — |
| Business Agreement Generator | 46 · Use with care | ★ 1.8k | 7mo ago | — |
| Contract Comparison Skill | 45 · Not recommended | ★ 1.8k | 7mo ago | — |
| Missing Protections Finder (AI Legal Assistant) | 45 · Not recommended | ★ 1.8k | 7mo ago | — |
| Counter-Proposal Generator | 41 · Not recommended | ★ 1.8k | 7mo ago | — |
How did FollowSkills review this skill?
The skill only reads user-specified files/URLs and writes one Markdown output file — narrow scope, no destructive defaults, an explicit legal disclaimer embedded in output, and largely transparent data flow. Deducted for: no user confirmation before writing output files, URL fetching that ingests external content without security warnings, unknown license metadata, unverified publisher identity, and incomplete confirmation/rollback practices versus the full least-privilege anchor.
Instructions are internally consistent with clear steps (read → translate → flag → summarize → output) and three input paths (file/URL/inline) are all handled; failure is diagnosable from the prompts. Deducted for: no tests, no edge-case handling (empty files, non-contract text, unreachable URLs) and no defined failure feedback; static review caps this at 10.
Trigger is explicit (/legal plain <file>), the audience (non-lawyers) and scenario (understanding contracts) are clear, and output structure is fully specified with a mandate to honestly flag ambiguity. Deducted for: English-only output with no declared fit for Chinese contracts or non-English jurisdictions; no explicit non-fit boundaries; dependence on Anthropic services raises mainland-China reachability concerns that the docs do not address.
The SKILL.md is well layered (trigger → steps → output template → guidelines) and the README covers install/uninstall and dependencies. Deducted for: no version number, changelog, or known-limitations disclosure for this skill; repository license metadata unknown; maintenance responsibility and update path unclear; marketing funneling to a Skool community.
A prompt-only skill whose design (clause-by-clause translation, five flag types, glossary, summary) plausibly delivers real marginal value at near-zero cost. Deducted for: no verified representative outputs, so translation accuracy and flag-judgment quality are unproven, and users must still self-verify legal meaning.
Only author claims and illustrative templates in source files; no test suite, CI evidence, third-party execution, or real output samples. README statistics (82%, 67%, etc.) are unsourced marketing claims. Auditable primary material is limited to the instructions themselves, with thin coverage.
Open a dimension to read why it scored that way
Evidence confidence：Low — Mostly static review, author material or a limited demo; useful for discovery, not high-risk decisions.
See the full review method →