agentsclimarketplace

Stilta patent drafter

Skill riteshkew/yc-skills/skills/stilta-patent-drafter

Catalog of all 198 YC Winter 2026 companies + 18 working Claude Code Skills inspired by the ones an open-source tool can replicate. Inspired by, not affiliated with.

Install
npx -y skills add riteshkew/yc-skills --skill stilta-patent-drafter

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Given an invention disclosure, retrieve related prior-art snippets from a local corpus via BM25 and draft a structured patent Background and Claims scaffold grounded in the disclosure and retrieved references.

SKILL.md

3.6 KB, 718 tokens by cl100k_base, as published. Nobody here has run it

Workflow

When this skill triggers, follow these steps in order.

Step 1 — Receive the Invention Disclosure

Accept a plain-text or Markdown invention disclosure from the user. The disclosure should describe:

  • The problem being solved
  • The novel technical features of the invention
  • Key components or steps of the system or method

If the user has not provided a disclosure, ask:

"Please paste or upload your invention disclosure. Describe the problem, the novel features of your invention, and the key components or steps involved."

Save the disclosure to a temporary file, e.g. disclosure.md.

Step 2 — Retrieve Prior Art via BM25

Run the dependency-free BM25 retriever over the committed prior-art corpus:

node scripts/retrieve.mjs resources/prior-art/ disclosure.md 3

The retriever:

  • Tokenizes the disclosure and each corpus document
  • Scores documents using the BM25 formula (k1=1.5, b=0.75)
  • Returns the top-K files with scores and matched terms as JSON

Capture the JSON output. If the process exits non-zero, report the stderr message to the user and stop.

Inspect the top results. For each, note:

  • The filename (maps to a prior-art reference)
  • The BM25 score (higher = more relevant)
  • The matched terms (shows which disclosure concepts overlap with prior art)

Step 3 — Draft Background and Claims

Run the scaffold assembler with the disclosure and retrieval results:

node scripts/draft.mjs disclosure.md retrieval.json draft.md

The assembler produces a Markdown file containing:

  1. Background — situates the invention against the retrieved prior art, citing each reference with its BM25 score and matched terms, followed by a verbatim excerpt of the disclosure.
  2. Claim 1 (Independent) — a system claim listing the key features extracted from the disclosure, in proper patent-claim format ("A system comprising… wherein…").
  3. Claim 2 (Dependent) — a dependent claim referencing claim 1 and adding a communication interface feature.

Present the drafted sections to the user for review.

Step 4 — Flag Novelty Risks

Review the matched terms between the disclosure and the top retrieved prior art. For each prior art reference:

  • Identify which disclosure features are already described in the prior art (potential novelty risks)
  • Highlight features that appear only in the disclosure and not in any retrieved reference (potential novel aspects)
  • Recommend that the user consult a registered patent attorney to evaluate patentability

Use this format:

Prior Art: [PA-1] <filename>
  Overlapping features: <list matched terms>
  Novelty risk: <brief assessment>

Novel aspects not found in retrieved prior art:
  - <feature A>
  - <feature B>

Example

See examples/input.md for a worked invention disclosure (smart umbrella with on-device rain prediction) and examples/output.md for the corresponding Background, Claims, and retrieval detail produced by the engine.

Run the example yourself from the skill root:

cd skills/stilta-patent-drafter
bash examples/run.sh

Important Disclaimer

This skill produces a structural scaffold for review purposes only. It is not legal advice and does not constitute a filed patent application. Patent claims require review and refinement by a registered patent attorney or agent before filing with any patent office.

What ships with it: 12 files

34.4 KB alongside SKILL.md, 3 of them executable

.claude-plugin/

examples/

scripts/

Gives 0 of the 12 instructions most legal skills give in 718 tokens

Counted across 234 of the 234 authors here whose files we hold, read 2026-08-07

  • Use text operators for text fieldsin 11 of 234, across 6 files
  • Consult qualified counsel before usein 11 of 234, across 3 files
  • Use PatentSearch API for patent searchesin 10 of 234, across 5 files
  • Confirm jurisdiction, employment type, and required clausesin 9 of 234, across 2 files
  • Choose a document template and tailor role-specific termsin 9 of 234, across 2 files
  • Validate compensation, benefits, and compliance requirementsin 9 of 234, across 2 files
  • Add signature, confidentiality, and IP assignment terms as neededin 9 of 234, across 2 files
  • Open the implementation playbook for detailed templatesin 9 of 234, across 2 files
  • Use TSDR for trademark data retrievalin 9 of 234, across 4 files
  • Ask for clarification if required inputs are missingin 8 of 234, across 2 files
  • Set the USPTO_API_KEY environment variablein 8 of 234, across 3 files
  • Use the uspto-opendata-python library for PEDSin 8 of 234, across 3 files

Said here and by no other author read

  • save the disclosure to a file
  • run the bm25 retriever over the prior-art corpus
  • stop and report if retriever exits non-zero
  • run the scaffold assembler with disclosure and retrieval results
  • present drafted sections to the user
  • identify overlapping disclosure features in prior art

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.