agentleFS
flowleap-aiflowleap-ai/flowleap-cli/skills/recipe-freedom-to-operate/SKILL.md

recipe-freedom-to-operate

Skill1 starsChanged 4 months agoMITView on GitHub

Freedom-to-operate clearance for a product or technology — per-feature query generation, dual-database blocking-patent search, and legal-status, family, and all-elements claim checks on every live candidate. Trigger when the user asks whether a product or technology can launch without infringing, or requests an FTO/clearance search.

The file

We split this skill into the rules it gives an agent

5 rules. About 0 are common, 0 are distinctive, and 5 are 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.

  • 5

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Search for potentially blocking patents for a product or technology.

    line 11
  2. Very specific:

    Write a query per key product feature, in each database's own syntax, yourself (method in flowleap-patent.

    line 21
  3. Very specific:

    A missing patent-data key never narrows a clearance scope: If one office answers provider_keys_required, search the live office fully and …

    line 43
  4. Very specific:

    Read the jurisdiction from the right field.

    line 70
  5. Very specific:

    If the full skill pack is installed, continue with recipe-infringement-charting to chart the at-risk claims element by element.

    line 124

≈ “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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Recipe: Freedom-to-Operate (FTO) Search

Search for potentially blocking patents for a product or technology. Each database uses its own query syntax — see flowleap-uspto for the USPTO Lucene grammar.

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.

Steps

Step 1: Write Targeted Searches

Write a query per key product feature, in each database's own syntax, yourself (method in flowleap-patent; ODP Lucene in flowleap-uspto). For each feature: extract its specific noun phrases, build the CQL around the discriminating ones, and probe the count — FTO is a recall search, so refine toward broad (drop classifications, OR in synonyms) rather than precise:

# EPO CQL count probe for feature 1 (read `total` from the payload)
flowleap --json tools run search_patents query='<CQL for feature 1>' range=1-1 details=false
# Repeat for each feature; USPTO Lucene queries are title + metadata only

Done when every key feature has a probed EPO CQL query and a USPTO ODP query.

Step 2: Search for Potentially Blocking Patents

flowleap --json patent search --query "<CQL for feature 1>" --limit 20
flowleap --json uspto search --query "<recommended_query for feature 1>" --limit 20
# Repeat for each feature

A missing patent-data key never narrows a clearance scope. If one office answers provider_keys_required, search the live office fully and carry the gated office into the deliverable as an explicit missing-key gap — an unsearched jurisdiction is not a cleared one, and web-scraped patent data is not an acceptable substitute for it. Ask for the free key at the end of the turn; when the user adds it, re-run only that office and merge. See flowleap-keys.

Done when both databases have been searched for every feature, or the gated office is recorded as an open missing-key gap.

Step 3: Screen to Live, In-Market Candidates

Take the one-call snapshot (bibliography + legal status + family + term) and keep only patents that are legally alive and cover a jurisdiction you sell in:

flowleap --json summary <patent-number>

Or, if you need the pieces separately:

flowleap --json ops legal <patent-number>     # is it active? which states?
flowleap --json ops family <patent-number>    # where is it filed?

Read the jurisdiction from the right field. family names the offices the invention published in; for an EP member that is one entry ("EP") and says nothing about countries. The countries are the designatedStates on the legal status (extensionStates for extension/validation states) — for a European regional filing those ARE its designated-state coverage. Then subtract the states lapsed per the PG25/PGFP events, which each carry their own state: a patent designating DE and still paying fees there blocks you in Germany, one lapsed there does not.

The designation is rolled up from the AK event with the latest date. If that event carried no readable list an older one answers, so the set can be a superseded, wider one — it can name a state the applicant has since dropped. For a clearance call that is the dangerous direction, so read the AK rows in events before you flag or clear a market on the roll-up alone.

Done when every survivor is both in force and designates a sales jurisdiction; drop the rest.

Step 4: All-Elements Claim Mapping

For each surviving blocker, pull its claims and map product features against every element of each independent claim:

flowleap --json ops claims <patent-number>

For an office with no text source, use the family-member workaround in flowleap-shared.

An unread claim is never a cleared claim. A survivor at an office with no text source answers FULLTEXT_UNAVAILABLE. When the workaround finds a served member, chart from that member's claims and label the chart "member text, not the office's claims": a reading aid, not a verdict on the office's right. When no member's text is served, carry the survivor into the deliverable as "claims not readable here", an open gap on the same footing as a missing-key gap: never cleared, never dropped.

A claim clears only if at least one of its elements is absent from the product. Done when every independent claim of every survivor is either cleared (≥1 element absent), flagged as a live infringement risk, or recorded as "claims not readable here".

Output

FTO data package per live blocking patent:

  • Legal status (active, expired, abandoned) and remaining term
  • Geographic reach against your sales jurisdictions — family members for the offices, designatedStates minus lapsed states for an EP filing's countries
  • Per-claim all-elements verdict: cleared or at-risk; a chart built on a family member's claims carries the label "member text, not the office's claims"; a right whose claims no source serves is "claims not readable here", an open gap

If the full skill pack is installed, continue with recipe-infringement-charting to chart the at-risk claims element by element.

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.