Code Patent Validator

## Quick judgement This skill is well-designed for its niche: turning code-patent-scanner findings into structured, multi-source search strategies and evidence maps, while staying clearly away from legal conclusions. Within the OpenClaw ecosystem it fills a focused "validator/strategist" role rather than a scoring or legal-evaluation role. **Overall verdict:** Strong, safety-conscious research aid for IP-adjacent workflows, especially when paired with code-patent-scanner. Its main limitations are inherent to the problem domain (cannot guarantee completeness or correctness of search strategies) rather than to the skill design itself. --- ## Strengths - **Tight scope & clear boundaries:** Explicitly does not rescore patterns, does not perform searches, and does not opine on legal status. It repeatedly positions itself as a research helper only. - **Well-structured output:** Produces queries per source (Google Patents, USPTO, GitHub, Stack Overflow), search priority recommendations by pattern type, differentiation questions, and a detailed evidence_map keyed by claim angles. - **Good error handling:** Specifies responses for empty input, invalid JSON, missing fields, and low-score patterns. This increases robustness in realistic user workflows. - **Workflow fit:** Designed to consume `patterns.json` from code-patent-scanner and emit `search_strategies.json`, with a clear recommended workflow ending in consultation with an attorney. - **Safety in terminology:** Enforces replacement of sensitive legal terms (e.g., avoids “patentable”, “non-obvious”) with more neutral language like “distinctive” or “existing implementations.” --- ## Key risks & limitations 1. **User over-interpretation risk:** Even with disclaimers, some users may treat generated queries or differentiation prompts as implicit legal signaling. The skill mitigates this via strong disclaimers but cannot fully prevent misinterpretation. 2. **Coverage and accuracy of queries:** The quality of research depends heavily on how well the generated queries capture the user’s pattern. There is no automatic feedback loop from real search results, so irrelevant or incomplete query sets are possible. 3. **Dependence on upstream scanner quality:** It relies on code-patent-scanner’s scores and pattern extraction. If the scanner is weak or misconfigured, the validator may build detailed strategies around low-value or mischaracterized patterns. 4. **Structured-input fragility:** While there is explicit error handling, the best behavior assumes patterns.json matches the expected structure. Deviations or partially structured data could lead to skipped patterns or degraded guidance. 5. **Confidentiality considerations (deployment-dependent):** In real use, users may feed sensitive code descriptions into the skill. Actual privacy/safety outcomes depend on the hosting environment rather than the prompt design itself. --- ## Recommended scenarios Use this skill when: - You have **already run code-patent-scanner** and want to systematically research whether similar implementations appear in public codebases, Q&A sites, or patent databases. - A technical founder, engineer, or IP-aware product manager wants to **prepare for a meeting with a patent attorney** by organizing evidence (files, commits, design docs, metrics) and pre-running searches. - An engineering team wants consistent **evidence mapping and differentiation prompts** for candidate patterns across a larger system. - You need a structured, repeatable way to create **search strategies for multiple patterns** rather than ad-hoc, one-off searches. It is less suitable as a standalone tool for users who have not yet analyzed their code (no scanner output) or who are expecting any form of legal assessment or automated database search.