From agent-almanac
Searches patents, academic papers, products, and open source for prior art to assess novelty/obviousness, challenge validity, support FTO, or document defensive publications.
npx claudepluginhub pjt222/agent-almanacThis skill uses the workspace's default tool permissions.
---
Guides structured prior art searches using Espacenet: CQL query construction, abstract triage, full-text keyword search, paginated claim reading, family expansion, report generation. For patent novelty and invalidation.
Provides IP guidance for developers: prior art searches, patentability assessments, claim drafting, strategy advice, full patent drafts, and FTO analysis. Informational only.
Conducts 7-step prior art searches and patentability assessments using BigQuery keyword/CPC queries, USPTO API, and timeline analysis for invention evaluation.
Share bugs, ideas, or general feedback.
Conduct a structured prior art search to find publications, patents, products, or disclosures that predate a specific invention. Used to assess patentability (can this be patented?), challenge validity (should this patent have been granted?), or establish freedom-to-operate (is this design covered by existing rights?).
Break the invention into its constituent technical features.
Search Map Example:
+------------------+-----------------------------------+-----------+
| Element | Search Terms | Priority |
+------------------+-----------------------------------+-----------+
| Attention layer | attention mechanism, self- | High |
| | attention, multi-head attention | |
| Sparse routing | mixture of experts, sparse MoE, | High |
| | top-k routing, expert selection | |
| Training method | knowledge distillation, teacher- | Medium |
| | student, progressive training | |
+------------------+-----------------------------------+-----------+
Expected: A complete decomposition with search terms for each element. The novel combination is identified — this is what the search must either find (to invalidate) or confirm is absent (to support novelty).
On failure: If the invention is too abstract to decompose, ask for a more specific description. If the claims are unclear, focus on the broadest reasonable interpretation of each claim element.
Search patent databases systematically.
Expected: A classified list of patent references mapped to the invention's elements. X references (if found) are showstoppers for novelty. Y references are the building blocks for obviousness arguments.
On failure: If no relevant patent art is found, this doesn't mean the invention is novel — non-patent literature (Step 3) may contain the critical reference. Absence in one database doesn't mean absence everywhere.
Search academic papers, products, open source, and other non-patent disclosures.
Expected: Non-patent references that complement the patent search. Academic papers and open-source code are often the most powerful prior art because they tend to describe technical details more explicitly than patents.
On failure: If non-patent literature is sparse, the technology may be primarily developed in corporate R&D (patent-heavy). Shift emphasis to patent literature and focus on the combination-based obviousness argument.
Evaluate how the collected prior art relates to the invention.
Claim Element vs. Prior Art Matrix:
+------------------+--------+--------+--------+--------+
| Element | Ref #1 | Ref #2 | Ref #3 | Ref #4 |
+------------------+--------+--------+--------+--------+
| Element A | X | X | | X |
| Element B | | X | X | |
| Element C | X | | X | |
| Novel combo A+B+C| | | | |
+------------------+--------+--------+--------+--------+
X = element disclosed in this reference
Expected: A clear claim chart showing which elements are covered by which references, with an assessment of novelty and obviousness. Each mapping cites specific passages or figures in the references.
On failure: If the claim chart shows gaps (elements not found in any prior art), those gaps represent the potentially novel aspects. Focus follow-up searches on those specific gaps.
Package the search results for their intended use.
Expected: A complete, well-organized search report that directly supports the intended decision. References are accessible and analysis is traceable.
On failure: If the search is inconclusive (no strong X or Y references, but some relevant background), state the conclusion clearly: "No anticipatory art found; closest art addresses elements A and B but not C. Recommend filing with claims emphasizing element C." Inconclusive is a valid and useful result.
assess-ip-landscape — Broader landscape mapping that contextualizes specific prior art searchesscreen-trademark — Trademark-specific conflict screening (different databases, different legal framework than patent prior art)file-trademark — Trademark filing procedures for when screening is completereview-research — Literature review methodology overlaps significantly with prior art searchsecurity-audit-codebase — Systematic search methodology parallels (thoroughness, documentation, reproducibility)