recipe-audit-report
Skill1 starsChanged 4 months agoMITView on GitHub
Governance recipe for producing an auditable record of AI-assisted patent research — reproducible command log, data provenance for every finding, portfolio status verification, and an AI-usage disclosure section suitable for internal or filing-adjacent records. Trigger when the user asks for an audit trail of patent research, an AI-assistance disclosure, or a verifiable methodology write-up of analysis work.
In this repository34 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- flowleap-academicSkill
- flowleap-authSkill
- flowleap-citationSkill
- flowleap-keysSkill
- flowleap-legalSkill
- flowleap-nplSkill
- flowleap-opsSkill
- flowleap-patentSkill
- flowleap-patstat-graphSkill
- flowleap-patstatSkill
- flowleap-sharedSkill
- flowleapSkill
- flowleap-toolsSkill
- flowleap-usptoSkill
- persona-ip-analystSkill
- persona-patent-attorneySkill
- persona-researcherSkill
- persona-startup-founderSkill
- recipe-academic-literature-reviewSkill
- recipe-claim-analysisSkill
- recipe-claim-draftingSkill
- recipe-custom-dashboardSkill
- recipe-find-betterSkill
- recipe-freedom-to-operateSkill
- recipe-infringement-chartingSkill
- recipe-invalidity-analysisSkill
- recipe-invention-disclosureSkill
- recipe-maintenance-feesSkill
- recipe-office-action-responseSkill
- recipe-patent-landscapeSkill
- recipe-patent-to-reportSkill
- recipe-prior-art-searchSkill
The file
We split this skill into the rules it gives an agent
1 rule. About 0 are common, 0 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 79
If the research being audited produced a recipe-custom-dashboard bundle, cite it rather than re-deriving its numbers: the bundle's own …
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Recipe: Audit Report (AI-Assisted Research)
Make AI-assisted patent research verifiable: every finding traceable to a command, every command re-runnable.
In a chat client with the FlowLeap connector or in the FlowLeap Patent AI app, call the tools named in the flowleap-shared tool table instead of these commands, and never run a flowleap command there. The steps stay the same.
Step 1: Record the Environment
flowleap --version
flowleap --json health api # backend apiVersion (the server build)
flowleap --json doctor # backend, auth mode, provider-key statusRecord: CLI version, backend apiVersion, backend base URL, date/time of the
session, and the auth mode (OAuth session vs. API token) — never token
values or key material.
Both versions matter for re-runnability. The CLI sends its own build as the
Client version header (X-FlowLeap-Client: cli/<version>), so a later
reader can tell which client produced the log; apiVersion pins the server side
of the same pair.
Step 2: Provenance Discipline During Research
For every substantive finding, log the exact command and keep the --json
output. The Tools facade makes provenance explicit — one surface, one envelope,
one named tool per capability, so "which system answered this" is never a guess:
flowleap --json tools list # capabilities available that day
flowleap --json tools run server_info # backend identity/versionNote per result: the source system (EPO OPS, USPTO ODP, OpenAlex, Semantic
Scholar, legal RAG), the query string, limits and date filters, and the
cached flag from the response envelope (cached data may lag live records).
Step 3: Verify Asset Status Claims
Any statement like "patent X is in force" must be backed by a dated check:
flowleap --json summary <patent-number> # legal status + family + term
flowleap --json timeline <patent-number> # event history behind the statusStep 4: Reproducibility Pass
Re-run the load-bearing commands at report time and diff against the original outputs. Patent registers move — note any finding that changed between research and reporting, with both dates.
Step 5: AI-Usage Disclosure Section
State plainly:
- Which steps were AI-assisted (search strategy, summarization, charting) and which tool/model performed them
- Which outputs a human verified, and how (e.g. claims read in full, legal-status events checked against the register)
- Known limitations: search recall is not exhaustive; OCR and summarization can introduce errors; results reflect database coverage on the query date
For filings, read the current duty-of-disclosure guidance from the office's
own page with flowleap-legal (37 CFR 1.56, MPEP 2001) and quote it with its
URL and edition.
Visual deliverable
If the research being audited produced a recipe-custom-dashboard bundle,
cite it rather than re-deriving its numbers: the bundle's own provenance
footer and reproduce block already meet this recipe's Verified-Data Contract
bar (sources, parameters, Data Edition where applicable, timestamps),
so Steps 1–4 can point at that bundle instead of re-collecting the same
evidence by hand.
Output
- Environment record (versions, backend, date, auth mode)
- Methodology section (sources, queries, filters — per finding)
- Status-verification table with check dates
- Command appendix: every command, in order, with output digests
- AI-usage disclosure with human-verification points and limitations
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

