Back to search
openapi-to-skills
SkillGitHub
OpenAPI to Agent Skill for context-efficient AI agents
Safety scan
A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.
Checked today
Badge for your README
Markdown
[](/listings/neutree-ai--openapi-to-skills)HTML
<a href="/listings/neutree-ai--openapi-to-skills"><img src="/api/listings/neutree-ai--openapi-to-skills/badge.svg" alt="findsafeskills safety rating" /></a>Install
i
What was scanned
The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.
README.md · README · 89 lines
View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)
This file has changed since it was scanned, so this is not exactly the text that was scanned, and no line is marked. The findings are listed with the safety scan above.
- # openapi-to-skills
- [](https://codecov.io/gh/neutree-ai/openapi-to-skills)
- [](https://www.npmjs.com/package/openapi-to-skills)
- Convert OpenAPI specifications into [Agent Skills](https://agentskills.io/) format - structured markdown documentation that minimizes context size for AI agents.
- AI agents often need domain-specific knowledge to complete tasks. Rather than writing custom prompts or building MCP tools from scratch, leveraging existing API documentation is a practical approach - and OpenAPI is the de facto standard with battle-tested specifications maintained over years.
- However, feeding raw OpenAPI specs to agents has limitations. Complex specifications can exceed LLM context limits, and even when they fit, loading the entire document for every request wastes valuable context.
- Agent Skills solves this by structuring documentation for on-demand reading. Agents load only what they need - starting with an overview, then drilling into specific operations or schemas. Since file reading is a universal capability across agent frameworks, this approach works everywhere without special integrations.
- ## Features
- - **Semantic structure** - Output organized by resources, operations, and schemas, enabling agents to load only relevant sections
- - **Smart grouping** - Operations grouped by tags or path prefix (auto-detected), schemas grouped by naming prefix
- - **Filtering** - Include/exclude by tags, paths, or deprecated status
- - **Customizable templates** - Override default Eta templates for custom output format
- ## Usage
- ```bash
- npx openapi-to-skills ./openapi.yaml -o ./output
- # or
- bunx openapi-to-skills ./openapi.yaml -o ./output
- # or from URL
- npx openapi-to-skills https://example.com/openapi.yaml -o ./output
- ```
- For agent-guided generation, use the [openapi-to-skills skill](./skills/openapi-to-skills/SKILL.md).
- See its [CLI reference](./skills/openapi-to-skills/references/cli.md) for all options, layouts, overlays, validation, and regeneration behavior.
- ## Programmatic API
- The package exports `convertOpenAPIToSkill` for integration into build pipelines or custom tooling:
- ```typescript
- import { convertOpenAPIToSkill } from 'openapi-to-skills';
- const spec = { /* OpenAPI spec object */ };
- await convertOpenAPIToSkill(spec, {
- outputDir: './output',
- parser: {
- skillName: 'my-api',
- filter: {
- includeTags: ['users', 'orders'],
- excludeDeprecated: true,
- },
- },
- });
- ```
- For advanced use cases, individual components (`createParser`, `createRenderer`, `createWriter`) are also exported.
- ## Examples
- ### Petstore (simple)
- See [examples/](./examples) for a simple demo with the classic Petstore spec (3 resources, 19 operations).
- ### Public APIs Collection (production-scale)
- [public-api-skills](https://github.com/Yuyz0112/public-api-skills) demonstrates real-world usage across a large collection of public APIs:
- | Metric | Original | Agent Skills |
- |--------|----------|--------------|
- | Source | 1 file (7.2 MB) | 2,135 files |
- | Resources | - | 77 index files |
- | Operations | 588 (all in one) | 588 individual files |
- | Schemas | 1,315 (all in one) | 1,468 individual files |
- A 7.2 MB monolithic spec becomes 2,135 focused files that agents can navigate on-demand.
- ## Roadmap
- - [ ] Detect unresolved external `$ref` and warn users
- - [x] Support fetching OpenAPI specs from URL
- ## Contributing
- Development guidelines are maintained in [CLAUDE.md](./CLAUDE.md), designed as a shared reference for both human contributors and AI agents.
- We welcome AI-assisted development. However, whether a patch is written manually or with AI collaboration, the author must fully review all changes before submission and take responsibility for the content.
- ## License
- Apache-2.0
Ratings and comments
Sign in to rate this skill.
Sign inFree to use. If it’s helped, pass it on.