Review US Patent Claims

Review US patent claims

Purpose and boundary

Provide evidence-based US patent drafting and examination-quality assistance. Do not claim to be a licensed attorney or patent agent, and do not present the output as legal advice, a filing clearance, a validity opinion, an infringement opinion, or a guarantee of allowance.

Use references/review-checklist.md for every full or complex review.

Current USPTO examination guidance and case law can change. Re-check the relevant official authority for the actual filing date, application route, procedural posture, technology, and review date.

Supported materials

• pasted claims;
• DOCX, PDF, TXT, or Markdown claim sets;
• complete application/specification;
• drawings and sequence listings;
• Office actions and draft responses;
• claim amendments and redlines;
• PCT applications intended for US national-stage or continuation practice; and
• identified published US applications/patents.

If a file is unreadable, request an accessible copy or pasted text. Do not infer missing text.

Intake

Capture:

• exact document and claim version;
• source filename and date;
• application type: provisional, nonprovisional under 35 U.S.C. 111(a), national stage under 35 U.S.C. 371, continuation, divisional, CIP, reissue, or unknown;
• application/publication/patent number if available;
• priority and filing dates;
• pre-AIA/AIA status where relevant;
• prosecution stage and deadline;
• current claims, marked-up claims, and clean claims;
• specification, abstract, drawings, examples, sequence listing, deposit/data, and prior-art discussion supplied;
• Office-action grounds and cited references;
• target commercial embodiment and likely infringing actor;
• desired breadth and fallback positions;
• jurisdictions from which language was adapted;
• whether filing already occurred; and
• whether claim amendments may add subject matter.

For AI-assisted inventions, flag inventorship facts for qualified review under current USPTO guidance. Do not decide inventorship from prompt history alone.

Document-availability matrix

Before analysis, state:

If only claims are supplied, label written description, enablement, corresponding structure/algorithm, and amendment basis as preliminary.

Optional PatSnap connectors

No connector is required for supplied authoritative materials.

Patent Briefing — recommended for identified publications

• https://open.patsnap.com/marketplace/mcp-servers/patent-briefing
• key: patent_briefing
• Official marketplace page: https://open.patsnap.com/marketplace/mcp-servers/patent-briefing

Use to retrieve or cross-check bibliography, family, status, claims, description, translations, and images. Verify against the authoritative application/prosecution record when material.

Advanced Patent Search — recommended only for authorized search

• https://open.patsnap.com/marketplace/mcp-servers/patent-search
• key: advanced_patent_search
• Official marketplace page: https://open.patsnap.com/marketplace/mcp-servers/patent-search

Use to resolve identifiers or conduct an authorized prior-art search. Record exact query, fields, dates, authorities, family rule, references reviewed, and claim-feature mapping.

Do not call a structural 102/103 review a novelty search.

Current authority orientation

Use official current USPTO/MPEP material as examination-practice guidance and distinguish it from statutes, regulations, precedential case law, and litigation claim construction.

Official pages checked for migration on 2026-08-07:

• Subject-matter eligibility: https://www.uspto.gov/patents/laws/examination-policy/subject-matter-eligibility
• MPEP 2106: https://www.uspto.gov/web/offices/pac/mpep/s2106.html
• BRI, MPEP 2111: https://www.uspto.gov/web/offices/pac/mpep/s2111.html
• 112(f), MPEP 2181: https://www.uspto.gov/web/offices/pac/mpep/s2181.html
• Multiple dependent claims, MPEP 608.01(n): https://www.uspto.gov/web/offices/pac/mpep/s608.html
• Restriction, MPEP 802: https://www.uspto.gov/web/offices/pac/mpep/s802.html
• National restriction versus PCT/371 unity, MPEP 823: https://www.uspto.gov/web/offices/pac/mpep/s823.html

The current USPTO eligibility page identifies MPEP 2103–2106.07 as current guidance and lists later updates, including the 2024 AI eligibility update, 2025 reminders/Desjardins-related changes, and 2026 Rule 132 declaration best practices. Check the page again in a live matter.

Workflow

1. Identify scope and authority

• Confirm whether review covers claims only, full application, amendment, Office-action response, or portfolio strategy.
• Identify application route and procedural posture.
• Identify review date and current authority cutoff.
• Separate pre-filing drafting suggestions from post-filing amendments.
• Identify deadlines but do not calculate a dispositive deadline without authoritative docket information.

2. Extract the claim set

Preserve:

• claim number;
• exact text;
• status if known (originally filed, currently pending, canceled, withdrawn, allowed, rejected, amended);
• independent/dependent/multiple-dependent form;
• statutory category;
• parent reference(s);
• dependency path;
• amendment changes; and
• source page/paragraph/locator.

Do not silently correct text before logging the issue.

3. Build the dependency graph

Check:

• duplicate or missing claim numbers;
• reference to nonexistent/canceled claims;
• circular or impossible dependencies;
• dependent claims that fail to further limit the parent;
• multiple dependent claims referring in the alternative only;
• a multiple dependent claim depending from another multiple dependent claim;
• mixed-category dependencies;
• inconsistent incorporation of limitations; and
• fee/claim-count implications requiring current verification.

4. Identify claimed inventive concepts

For each independent claim:

• statutory category;
• actor/system boundary;
• essential elements/steps;
• apparent technical problem;
• claimed technical mechanism;
• result/effect;
• likely commercial embodiment;
• likely distinguishing limitations;
• evidence/source locator; and
• uncertainty.

Do not substitute the abstract or marketing description for the claim language.

5. Apply BRI for pending US examination

MPEP 2111 states that pending claims receive the broadest reasonable interpretation consistent with the specification, not the broadest possible interpretation.

Review:

• ordinary/customary meaning to a person of ordinary skill;
• express definitions;
• consistent specification/drawing usage;
• preamble effect;
• transitional phrase;
• functional language;
• intended use;
• wherein/whereby clauses;
• relative terms;
• ranges; and
• open/closed language.

Do not apply prosecution BRI as though it were the issued-claim litigation standard. State the procedural context.

6. Review 35 U.S.C. 112(a)

Written description

For each limitation, identify support by paragraph/page/figure/example/sequence.

Review:

• genus/species breadth;
• ranges and endpoints;
• alternatives/Markush groups;
• combinations and subcombinations;
• negative limitations;
• optional features;
• functional results;
• newly introduced terminology;
• priority entitlement concerns; and
• new-matter risk for amendments.

Enablement

Review the full claimed scope and technology-specific predictability, working examples, guidance, variability, testing burden, and undue experimentation risk under current law.

For software/AI, inspect disclosed architecture, data flow, model/training/inference steps, control logic, technical improvement, and implementation detail.

For chemistry, biotech, pharma, diagnostics, and materials, inspect representative species, ranges, protocols, structure/function, endpoints, utility, and evidence across the claimed scope.

Do not conclude enablement from one example without analyzing scope.

Best mode

For applications subject to AIA treatment, distinguish the disclosure requirement from enforceability consequences; route matter-specific advice to qualified counsel.

7. Review 35 U.S.C. 112(b)

Check:

• antecedent basis;
• inconsistent terminology;
• unclear boundaries;
• subjective/relative terms;
• terms of degree and objective standards;
• optional/alternative language;
• unclear step order;
• actor ambiguity;
• mixed statutory categories;
• intended-use/result-only language;
• internally inconsistent ranges;
• claim-reference errors; and
• ambiguity introduced by translation.

Context controls the result. Words such as “about,” “substantially,” “module,” or “configured to” are not automatically indefinite.

8. Review 35 U.S.C. 112(f)

MPEP 2181 states application turns on claim language, not applicant intent.

For each function-associated term:

1. identify whether the claim uses “means” or “step”;
2. identify whether the term has sufficiently definite structural meaning in context;
3. identify the complete claimed function;
4. locate corresponding structure, material, or acts in the specification;
5. confirm linkage between disclosure and function;
6. for computer-implemented functions, locate a sufficient algorithm where required;
7. assess indefiniteness/scope consequences; and
8. compare structural recasting if 112(f) is not intended.

Do not treat a list of nonce words as a per se rule.

9. Review 35 U.S.C. 101

Use the current USPTO subject-matter-eligibility framework and current controlling law.

Analyze each claim as a whole:

• statutory category;
• judicial exception, if recited;
• Step 2A Prong One;
• Step 2A Prong Two/practical application;
• Step 2B/significantly more;
• claimed technological improvement;
• specification support for that improvement;
• additional elements and their arrangement; and
• claim-specific outcome.

For software/AI, do not use a fixed list of magic eligible terms. Evaluate concrete improvements to computer functionality or another technical field and the claim as a whole.

For diagnostics/life sciences, distinguish natural correlation/phenomenon/product concerns from supported concrete treatment, preparation, assay configuration, transformation, or other claim elements.

Do not conflate eligibility with novelty, obviousness, or disclosure.

10. Review 35 U.S.C. 102 and 103

Without an executed search

Perform only structural vulnerability review:

• conventional element aggregation;
• apparent predictable combination;
• routine optimization/range selection;
• result-only distinction;
• missing technical mechanism;
• weak nexus between feature and effect;
• likely single-reference exposure; and
• fallback limitation quality.

Label this search_status: not_executed.

With cited references or an executed search

For each ground/reference:

• verify publication date and prior-art eligibility for the relevant filing/priority context;
• map every claim limitation;
• identify explicit, implicit, or inherency theory;
• for 103, identify combination rationale, analogous art, reasonable expectation of success, teaching away, unexpected results, objective indicia and nexus where supported;
• distinguish examiner assertion from verified reference disclosure;
• record missing limitation/evidence; and
• avoid final novelty/nonobviousness conclusions without qualified review.

11. Review claim architecture

Assess:

• independent claim categories;
• commercial actor/product coverage;
• broad, intermediate, and narrow fallback layers;
• meaningful dependent claims;
• parallel system/method/CRM/apparatus/composition/manufacture/kit claims as appropriate;
• unnecessary environment/UI/user/parameter limitations;
• product-by-process and intended-use issues;
• unsupported breadth;
• single-point-of-failure limitations;
• continuation/divisional/CIP options; and
• double-patenting/terminal-disclaimer issues for qualified review.

Do not add a claim category merely to increase count.

12. Review restriction/election

For US national applications under 111(a), review current MPEP Chapter 800 principles for independent/distinct inventions, species, product/process groups, burden and linking claims.

For 371 national stage/PCT context, distinguish unity-of-invention practice under MPEP 823/Chapter 1800.

Identify:

• invention groups;
• species/genus structure;
• design/operation/effect relationship;
• linking/generic claims;
• rejoinder potential;
• election consequences;
• divisional timing/35 U.S.C. 121 safe-harbor issues requiring counsel; and
• continuation strategy.

Do not predict a restriction requirement solely from multiple independent claims.

13. Review infringement observability and enforceability strategy

This is drafting strategy, not an infringement opinion.

Review:

• whether each limitation is externally observable;
• public documentation, testing, teardown, source-code discovery, network evidence, or manufacturing proof needed;
• single versus multiple actors;
• method-step attribution and divided-infringement risk;
• location/jurisdiction of system components or steps;
• product/system alternatives to hidden user behavior;
• easy design-arounds;
• claim differentiation and equivalents/estoppel sensitivity; and
• whether amendment language could create avoidable disclaimer.

14. Draft issue records

Each issue must include:

issue_id: US-CLM-001
severity: High|Medium|Low
category: BRI|112a|112b|112f|101|102|103|form|architecture|restriction|observability|strategy
claim_numbers: []
quoted_language: ""
claim_source_locator: ""
specification_support_locator: ""
authority_or_practice_basis: ""
risk: ""
confidence: High|Moderate|Limited
recommended_action: ""
example_language: ""
scope_tradeoff: ""
new_matter_caution: ""

Do not assign High merely because a keyword appears.

15. Severity guide

• High: likely material rejection/indefiniteness/support problem, missing core claim coverage, serious actor/observability failure, material dependency/form defect, or amendment with substantial new-matter risk.
• Medium: likely prosecution friction, 103 vulnerability, weak fallback, 112(f) ambiguity, restriction/election exposure, or avoidable strategic limitation.
• Low: clarity polish, formatting, optional architecture improvement, or minor practice preference.

State uncertainty and procedural context.

16. Amendment and redline strategy

For each proposed change:

• exact claim and text;
• objective;
• support locator;
• anticipated effect on BRI/101/102/103/112;
• scope gained/lost;
• actor/observability effect;
• dependency changes;
• new-matter warning;
• prosecution-history/disclaimer concern; and
• whether a dependent claim is preferable.

Post-filing additions must not introduce new matter. If support cannot be located, do not present language as safely amendable.

Output structure

# US patent claims review

## Overall assessment
[Preliminary risk, document completeness, prosecution posture, and readiness boundary.]

## Scope and materials
[Document/version matrix, authority cutoff, search status.]

## Claim-set map
[Independent claims, categories, dependencies, invention groups.]

## Priority issues
| Severity | Issue ID | Claim | Category | Quoted language | Risk | Support/authority | Recommended action | Confidence |
|---|---|---|---|---|---|---|---|---|

## Claim-by-claim review
### Claim [n]
- Scope/BRI:
- Category/dependency:
- 101:
- 102/103 search status and vulnerability:
- 112(a)/(b)/(f):
- Observability/actor:
- Recommendation/support:

## Amendment strategy
[Independent claims, dependent fallback ladder, continuation/divisional options, change log.]

## Example amendment text
[Only supported representative language, with additions/deletions clearly marked.]

## Limitations and attorney review gates
[Missing materials, search, current law, deadlines, new matter, claim construction, business facts.]

Final checks

• Exact claim version identified.
• Application route and stage identified.
• Claim numbers/categories/dependencies reconcile.
• BRI is prosecution-context and specification-consistent.
• 112(a) statements have support locators or are preliminary.
• 112(f) analysis identifies function and corresponding disclosure.
• 101 uses current claim-as-a-whole guidance.
• 102/103 search status is explicit.
• Multiple-dependent format follows current US rules.
• Restriction and PCT unity are distinguished.
• Proposed amendments have basis and new-matter review.
• Observability and actor issues are separate from infringement conclusions.
• Current official authorities are rechecked.
• Output is drafting/review assistance, not legal advice.