clause-extractor

# Clause Extractor ## Quick Start Extract, categorize, and compare specific clauses from legal documents. Useful for building clause libraries, comparing terms across vendors, or identifying standard vs. non-standard provisions. ## Workflow 1. Identify the target clauses: termination, liability, payment, confidentiality, IP, indemnification, SLA, etc. 2. Read the document(s) provided 3. Locate and extract each target clause with its exact wording 4. Categorize by clause type and label as standard/non-standard/aggressive 5. If comparing multiple documents, create a side-by-side comparison 6. Summarize findings with recommendations ## Examples **Example 1: Extract Key Clauses** Input: "Extract all payment and termination clauses from this vendor contract" Output: ``` CLAUSE EXTRACTION ================= Document: Vendor Service Agreement — ABC Corp Date Extracted: 03/23/2026 PAYMENT CLAUSES --------------- Clause 4.1 — Payment Terms: "Client shall pay all invoices within thirty (30) days of receipt. Late payments shall accrue interest at 1.5% per month." 📋 Category: Payment Terms 🏷️ Assessment: Standard — 30-day net is typical Clause 4.3 — Price Adjustment: "Provider reserves the right to adjust pricing with sixty (60) days written notice. Continued use constitutes acceptance of new pricing." 📋 Category: Price Escalation 🏷️ Assessment: Non-standard — no cap on increases, acceptance by inaction TERMINATION CLAUSES ------------------- Clause 8.1 — Termination for Convenience: "Either party may terminate with ninety (90) days written notice." 📋 Category: Convenience Termination 🏷️ Assessment: Standard — mutual right, reasonable notice period Clause 8.3 — Effect of Termination: "Upon termination, Client shall pay all outstanding fees and return all Provider materials. Provider shall retain all data for 90 days." 📋 Category: Post-Termination 🏷️ Assessment: Review needed — data retention clause may conflict with data protection requirements SUMMARY ------- Total clauses extracted: 4 Standard: 2 | Non-standard: 1 | Needs review: 1 ⚠️ Flag: Clause 4.3 allows unlimited price increases with passive acceptance. Recommend negotiating a price cap or active consent requirement. ``` **Example 2: Multi-Contract Comparison** Input: "Compare the SLA clauses across these 3 vendor proposals" Output: ``` CLAUSE COMPARISON ================= Clause Type: Service Level Agreement (SLA) | Term | Vendor A | Vendor B | Vendor C | |----------------|----------------|----------------|----------------| | Uptime | 99.9% | 99.5% | 99.95% | | Measurement | Monthly | Quarterly | Monthly | | Credits | 10% per 0.1% | 5% per 0.5% | 15% per 0.1% | | Max Credit | 30% monthly | 10% quarterly | 100% monthly | | Exclusions | Maintenance | Maintenance+3rd| Maintenance | | Response Time | 1hr critical | 4hr critical | 30min critical | RECOMMENDATION: Vendor C offers the strongest SLA terms with highest uptime guarantee and most generous credit structure. ``` ## Tools - Use `Read` to load contracts and legal documents - Use `Grep` to search for specific clause keywords across documents - Use `Write` to save extracted clauses or comparison reports - Use `Glob` to find multiple contract files in a directory ## Error Handling - If clause type not specified → ask which clauses to extract or offer common categories - If clause is ambiguous or poorly defined → extract as-is and flag for review - If document format prevents clean extraction → note formatting issues - If clause references external documents → note the dependency ## Connectors (Optional) This skill works standalone. When connected to external tools, it unlocks additional capabilities: | Connector | What it enables | |-----------|----------------| | ~~document management | Access contracts and legal documents from centralized storage | | ~~contract system | Pull clause data from contract lifecycle management platforms | | ~~email | Extract clauses from contracts shared as email attachments | | ~~compliance database | Tag extracted clauses against regulatory compliance requirements | ## Rules - Always preserve the exact original wording when extracting clauses - Label each clause with: section reference, category, and standard/non-standard assessment - For comparisons, use tabular format for easy scanning - Flag any clause that deviates significantly from industry standards - Note cross-references between clauses (e.g., termination triggers payment obligations) - Include the document source and extraction date for traceability - Never modify or paraphrase clause language without clearly marking it as a summary ## Output Template ``` CLAUSE EXTRACTION ================= Document: [Document title] Date Extracted: [MM/DD/YYYY] [CLAUSE TYPE] ------------- Clause [X.X] — [Clause Title]: "[Exact clause text]" 📋 Category: [Category] 🏷️ Assessment: [Standard / Non-standard / Aggressive] SUMMARY ------- Total clauses extracted: [N] Standard: [N] | Non-standard: [N] | Needs review: [N] ⚠️ [Key flags or recommendations] ``` ## Related Skills - `contract-reviewer` — For full contract reviews beyond individual clause extraction - `legal-summarizer` — For plain-language summaries of extracted clauses - `compliance-checker` — For checking extracted clauses against regulatory standards - `nda-drafter` — For drafting NDA clauses based on extracted best practices