Zhongzhi Xinjian Patent Office Action Response

Problem: Office Action Replies Tend to Be Generic

When responding to a patent office action, the hard part is not citing rules, but mapping each examiner finding—novelty, inventive step, clarity, support, unity, or amendment beyond scope—to specific paragraphs in the specification, cited references, and the Patent Examination Guidelines.

Workflow: Structured Decomposition and Argument Chains

The skill organizes the work into executable modules. It reads the office action, application files, and all cited references, then builds a defect checklist that records claim numbers, rejection grounds, evidence passages, and the examiner’s key language. It routes by OA type and selects a primary strategy around closest prior art, distinguishing features, technical suggestion, and routine means. Arguments use a fixed chain: premise, fact, reasoning, conclusion. For inventive step, it checks for fragmented feature evaluation, ignored synergy, and hindsight. For novelty, it applies the individual comparison rule. For formal defects, it prefers targeted claim amendments and verifies that the amendment remains within the original disclosure.

It can also use scripts/extract_figs.py to extract drawings and place side-by-side comparisons in the opinion statement, then run self-checks for full coverage, accurate citations, correct figure numbers, and strategy consistency. Outputs include an opinion statement, amended claims, drawing comparison materials, and a project log.

Boundaries

It does not cover new patent drafting or PCT international phase responses. Reexamination and third-party invalidation require their dedicated skills; this skill mainly reuses the inventive-step argumentation methodology.