pinecone-assistant-uspto — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited pinecone-assistant-uspto (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
This skill extends the base pinecone-assistant skill with USPTO-specific domain knowledge, query patterns, and multi-MCP integration workflows. The pinecone_assistant MCP serves as the knowledge layer (MPEP, guidelines, case law) that complements the data-access USPTO MCPs.
| MCP | Role | What It Provides |
|---|---|---|
| pinecone_assistant | Knowledge layer | MPEP guidance, examination standards, case law |
| PFW MCP | Data access | Prosecution history, file wrapper documents |
| PTAB MCP | Data access | IPR/PGR proceedings, board decisions |
| FPD MCP | Data access | First petition data, petition decisions |
| Citations MCP | Data access | Prior art citations, examiner citation patterns |
Always pair pinecone_assistant with the appropriate data MCP for complete research.
Match the patent law issue to the correct domain for assistant_strategic_multi_search_context or assistant_strategic_multi_search_chat:
| Legal Issue | Domain | When to Use |
|---|---|---|
| Alice/Mayo, abstract idea, software eligibility | section_101_eligibility | Any §101 rejection or eligibility question |
| KSR, Graham factors, motivation to combine | section_103_obviousness | Any §103 rejection or obviousness analysis |
| Written description, enablement, Wands factors | section_112_requirements | Any §112 rejection or specification question |
| Anticipation, single reference, inherent disclosure | section_102_novelty | Any §102 rejection or novelty analysis |
| Phillips standard, prosecution history estoppel | claim_construction | Claim scope disputes, litigation, licensing |
| IPR institution, BRI, PTAB estoppel | ptab_procedures | Post-grant proceedings, IPR/PGR analysis |
| Computer-implemented inventions, AI/ML patents | software_patents | Tech center 2100/2400 software/AI patents |
| Manufacturing, structural features, TC 3600/3700 | mechanical_patents | Mechanical invention examination |
| Broad patent law questions, unknown issue | general_patent_law | Starting point when issue is unclear |
Always use low temperature (0.0–0.3) for USPTO research with assistant_chat:
Step 1 — Research legal framework:
assistant_strategic_multi_search_context(
query="[invention description]",
domain="section_101_eligibility",
max_searches=3
)
→ Returns: Alice/Mayo two-step, technological improvement, inventive concept guidance
Step 2 — Pull office action and claims (PFW MCP):
pfw_get_application_documents(application_number="...", include_raw_xml=False)
pfw_get_document_by_id(...)
Step 3 — Targeted MPEP lookup for specific argument:
assistant_context(
query="MPEP 2106.05(a) technological improvement practical application",
top_k=4,
snippet_size=2048
)
Step 4 — Draft response with practical application argumentsStep 1 — Research legal framework:
assistant_strategic_multi_search_context(
query="[invention technology area]",
domain="section_103_obviousness",
max_searches=4
)
→ Returns: Graham factors, KSR 7 rationales, secondary considerations
Step 2 — Pull prior art citations (Citations MCP or PFW MCP):
Analyze what references the examiner cited
Step 3 — Targeted secondary considerations research:
assistant_context(
query="secondary considerations commercial success unexpected results long felt need",
top_k=4
)
Step 4 — Build Graham factors analysis and secondary considerations argumentsStep 1 — Identify type of §112 issue:
- §112(a): Written description / enablement
- §112(b): Indefiniteness
- §112(f): Means-plus-function
Step 2 — Research applicable standard:
assistant_strategic_multi_search_context(
query="[claim term or technology]",
domain="section_112_requirements",
max_searches=3
)
Step 3 — Targeted lookups by issue type:
§112(a): assistant_context(query="written description possession requirement MPEP 2163")
§112(b): assistant_context(query="Nautilus reasonable certainty indefiniteness MPEP 2173")
§112(f): assistant_context(query="means-plus-function corresponding structure MPEP 2181")
Step 4 — Pull specification via PFW MCP for comparisonStep 1 — Research institution standards:
assistant_strategic_multi_search_context(
query="IPR institution petition standards",
domain="ptab_procedures",
max_searches=3
)
Step 2 — Find existing PTAB proceedings (PTAB MCP):
search_trials_minimal(patent_number="...", limit=10)
Step 3 — Research relevant legal issues in the proceeding:
assistant_context(
query="[primary rejection type in IPR] PTAB institution decision",
top_k=5
)
Step 4 — Analyze claim construction approach:
assistant_strategic_multi_search_context(
query="[claim terms in dispute]",
domain="claim_construction"
)Step 1 — Detect primary vulnerability from prosecution history
Step 2 — Research legal framework via pinecone_assistant:
assistant_strategic_multi_search_context(domain="[detected domain]")
Step 3 — Pull prosecution history (PFW MCP)
Step 4 — Check PTAB history (PTAB MCP)
Step 5 — Analyze citations (Citations MCP)
Step 6 — Deep-dive research on secondary issues:
assistant_context(query="[specific legal issue]", top_k=5)
Step 7 — Synthesize defense strategyInclude statutory section numbers and MPEP references:
GOOD: "Section 101 Alice Mayo two-step framework abstract idea software"
GOOD: "MPEP 2106.04 judicial exceptions categories analysis"
GOOD: "KSR motivation to combine predictable results rationale"
GOOD: "Nautilus reasonable certainty indefiniteness §112(b) POSITA"Include technology specifics when relevant:
GOOD: "AI machine learning patent eligibility practical application §101"
GOOD: "pharmaceutical formulation obviousness teaching away secondary considerations"
GOOD: "computer-implemented method abstract idea technological improvement"Include case names for targeted case law research:
GOOD: "Alice Corp CLS Bank abstract idea step two significantly more"
GOOD: "KSR Teleflex motivation combine seven rationales"
GOOD: "Graham John Deere four factors secondary considerations"
GOOD: "Phillips AWH intrinsic extrinsic evidence claim construction"
GOOD: "Nautilus Biosig reasonable certainty definiteness""Section 101 Alice Mayo two-step abstract idea [technology]"
"MPEP 2106 judicial exceptions practical application"
"software patent eligibility technological improvement computer functionality"
"AI machine learning patent §101 significantly more inventive concept""KSR motivation to combine [technology] predictable results"
"Graham factors scope prior art level ordinary skill POSITA [field]"
"secondary considerations commercial success unexpected results [invention type]"
"teaching away prior art [technology] motivation combine""written description possession requirement MPEP 2163 [technology]"
"enablement Wands factors undue experimentation [technology]"
"indefiniteness Nautilus reasonable certainty [claim term]"
"means-plus-function §112(f) corresponding structure specification""Phillips claim construction intrinsic evidence specification [term]"
"prosecution history estoppel argument-based surrender [amendment type]"
"functional claiming 112(f) module mechanism structure disclosure"Maximize coverage across the MPEP with multiple specialized assistants:
mpep-core-assistant → MPEP 9th Edition core chapters (Parts 1-4)
mpep-updates-assistant → Federal Register notices and recent guidances
mpep-training-assistant → USPTO training materials and examination examples
case-law-assistant → Federal Circuit and PTAB decisions
examination-assistant → Art-unit specific examination guidelinesResearch flow across assistants:
1. Start with mpep-core-assistant for foundational MPEP guidance
2. update_configuration(assistant_name="mpep-updates-assistant")
→ Check for recent guidance updates on the same topic
3. update_configuration(assistant_name="case-law-assistant")
→ Research supporting case law and PTAB precedent
4. Synthesize findings from all three sourcespinecone_assistant → What does the law/MPEP say about this issue?
PFW MCP → What does the actual prosecution file show?
Combined → Draft office action response or prosecution strategypinecone_assistant → What are the legal standards for IPR institution?
PTAB MCP → What have actual PTAB boards decided on similar patents?
Combined → Assess institution likelihood and build petition/responseCitations MCP → What prior art did the examiner cite? (post Oct 2017)
pinecone_assistant → What does MPEP say about anticipation/obviousness with these references?
Combined → Build targeted prior art argumentsFPD MCP → What petitions were filed? What did the Director decide?
pinecone_assistant → What is the legal basis for the petition type?
Combined → Analyze petition strategy and precedentUSPTO research sessions often involve multiple MCPs. Manage tokens carefully:
| Step | Tool | ~Token Cost |
|---|---|---|
| MPEP framework research | assistant_strategic_multi_search_context | 10-20K context |
| Targeted MPEP lookup | assistant_context | 5-10K context |
| Deep-dive case law | assistant_context (top_k=5) | 8-12K context |
| Prosecution history (PFW) | pfw_get_application_documents(include_raw_xml=False) | 3-5K |
| Office action retrieval | pfw_get_document_by_id | 10-30K |
| PTAB trial search | search_trials_minimal | 2-5K |
Reserve 20% of context for final synthesis and output.
Use `include_raw_xml=False` in all PFW calls unless specific XML content is needed — saves 45-75K tokens per call.
domain: "section_101_eligibility" OR "software_patents"
queries: [
"Alice step 2A Prong 2 practical application [technology]",
"technological improvement computer functionality §101",
"MPEP 2106.05(a) improvement technology"
]
PFW: Pull office action, claims, specificationdomain: "section_103_obviousness"
queries: [
"KSR motivation to combine [tech area]",
"secondary considerations commercial success unexpected results",
"teaching away prior art [specific feature]"
]
Citations MCP: Identify examiner's cited references
PFW: Pull office actiondomain: "section_102_novelty" then "section_103_obviousness"
queries: [
"anticipation single reference [technology]",
"obviousness combination [technology]",
"[key claim features] prior art disclosure"
]
No PFW needed — pure knowledge base researchdomain: "ptab_procedures" then "section_103_obviousness" or "section_102_novelty"
queries: [
"IPR institution decision reasonable likelihood standard",
"BRI broadest reasonable interpretation PTAB",
"[grounds for challenge] MPEP guidance"
]
PTAB MCP: Search similar institution decisions
PFW MCP: Pull prosecution history for estoppel analysisdomain: "claim_construction"
queries: [
"Phillips intrinsic evidence specification [claim term]",
"prosecution history estoppel [amendment type]",
"functional claiming means-plus-function [claim language]"
]
PFW MCP: Pull prosecution history~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.