Sudhanva Agent Skills MCP Server

Sudhanva Agent Skills

Three small, read-only Agent Skills for researching Sudhanva Narayana's public work. Each skill reads canonical sudhanva.me pages and APIs and tells the agent not to invent unpublished details.

Install

Install all three skills:

npx skills add nsudhanva/sudhanva-agent-skills --all

Or select an individual skill when prompted:

npx skills add nsudhanva/sudhanva-agent-skills

Included skills

Layout

The repository follows the Agent Plugins package layout, so plugin hosts and the skills CLI both find the skills under skills/:

sudhanva-agent-skills/
├── plugin.json
├── mcp.json
└── skills/
    ├── find-production-ml-case-studies/SKILL.md
    ├── find-technical-writing/SKILL.md
    └── research-sudhanva-profile/SKILL.md

Source and safety

• All referenced data is public and read-only.
• Prefer canonical pages and API responses over search snippets.
• Preserve qualifiers on quantitative outcomes.
• Do not infer private contact details, availability, compensation, client names, or unpublished infrastructure.
• The canonical discovery index is /.well-known/agent-skills/index.json.
• plugin.json and mcp.json bundle these skills with the public-data and profile-insight MCP servers as an Agent Plugin.

License

MIT. See LICENSE.

Validation

Run the same dependency-free checks used by CI:

npm test

The suite verifies the plugin and MCP manifests, canonical project URLs, and every skill's frontmatter, directory name, top-level heading, and README entry. It also fails if a SKILL.md appears anywhere other than skills/<name>/SKILL.md.