agentleFS
flowleap-aiflowleap-ai/flowleap-cli/skills/recipe-patent-landscape/SKILL.md

recipe-patent-landscape

Skill1 starsChanged 4 months agoMITView on GitHub

Patent-landscape analysis for a technology area — scoped dual-database search, key-player identification, full-corpus filing analytics, and CPC-versus-year white-space detection. Trigger when the user asks to map a technology space, identify who patents in an area, or report filing trends and white space.

The file

We split this skill into the rules it gives an agent

4 rules. About 0 are common, 0 are distinctive, and 4 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.

  • 4

    Very specific

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

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

    Write the scoping queries yourself (method in flowleap-patent): extract the area's specific noun phrases, keep at least one discriminating …

    line 21
  2. Very specific:

    Verify any CPC code against the official scheme before scoping on it — flowleap patstat query on flowleap.cpc_scheme (see flowleap-patent).

    line 33
  3. Very specific:

    If one office answers provider_keys_required, its patent-data key is missing: map the live office in full, label the landscape as covering …

    line 47
  4. Very specific:

    When tallying players or counts from the search results, collapse to one entry per patent family so multi-jurisdiction filings are not …

    line 96

≈ “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: Patent Landscape Analysis

Map the patent landscape for a technology area, identifying key players, trends, and gaps. 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: Define Search Scope

Write the scoping queries yourself (method in flowleap-patent): extract the area's specific noun phrases, keep at least one discriminating term (a CPC class alone maps a continent, not a landscape), and probe the count — a landscape query legitimately lands broader than a novelty query, but a count in the tens of thousands still means the scope is the technology area, not the landscape:

# Count probe (read `total` from the payload), then adjust scope
flowleap --json tools run search_patents query='<self-written CQL>' range=1-1 details=false

Verify any CPC code against the official scheme before scoping on it — flowleap patstat query on flowleap.cpc_scheme (see flowleap-patent); group titles carry the specific technology, the 4-char class only the headline.

Done when you have a probed EPO CQL query and a USPTO ODP query for the area.

Step 2: Broad Patent Search

flowleap --json patent search --query "<CQL from step 1>" --limit 50
flowleap --json uspto search --query "<ODP Lucene from step 1>" --limit 50

If one office answers provider_keys_required, its patent-data key is missing: map the live office in full, label the landscape as covering that office only because of a missing key — never as the shape of the field — and ask for the free key at the end. flowleap patstat aggregates stay available keyless, but they are twice-yearly snapshot counts, not a substitute for live search. See flowleap-keys.

Done when both databases have returned their result sets, or the gated office is named as an open missing-key gap.

Step 3: Corpus Analytics

# Filing trends by year, country and CPC breakdowns, top assignees
flowleap --json analytics --keyword "<technology>" --date-from 2015-01-01
flowleap --json analytics --cpc <cpc-prefix> --country US --date-from 2020-01-01

Step 4: Identify Key Players

Scope the Step 1 query per applicant with pa= — wildcards catch name variants (pa=GOOGLE* catches "Google LLC" and "Google Inc"), and subsidiaries file separately (Google also files as Alphabet, DeepMind, Waymo):

flowleap --json patent search --query 'pa=<ASSIGNEE>* AND <discriminating terms from step 1>' --limit 30

Step 5: Check Recent Activity

patent search returns relevance-ranked hits; ops search --cql adds CQL date-range filtering (pd>=2024) for a time-sliced view the ranked search does not expose:

flowleap ops search --cql "ti=<technology> AND pd>=2024" --start 1 --end 50

Step 6: Flag White Space

Cross the analytics CPC breakdown against the filing-year trend to flag subclasses that are sparse or declining while neighbours grow. Done when at least one sparse/declining CPC subclass (or a confirmed absence) is identified.

Output

A dataset segmented by database, applicant, and filing date, plus corpus-level trend charts (filings per year, top assignees, CPC and country distributions). When tallying players or counts from the search results, collapse to one entry per patent family so multi-jurisdiction filings are not double-counted; the corpus analytics figures are aggregate backend counts, reported as returned.

Visual deliverable

To turn Step 3's filing-trend numbers and Step 6's white-space finding into a shareable HTML dashboard, follow this recipe's analysis through to the end, then render it with recipe-custom-dashboard — its landscape white-space template (CPC × year heatmap) is built for this recipe's final step, and its filing-trends template covers Step 3's year-over-year counts. Analysis logic stays here; the dashboard skill only owns presentation.

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.