A collection of agent skills for working with DeepSource products. Skills are packaged instructions and scripts that extend AI coding agents with DeepSource capabilities.
Scan repositories for security vulnerabilities, leaked secrets, and dependency issues using the Autofix Bot API, and auto-fix detected issues.
- Use when: you want to run a security scan on a repository or codebase using Autofix Bot
- Use when: you want to analyze code changes or pull requests for vulnerabilities
- Use when: you want to auto-fix security issues detected by Autofix Bot
- Includes: sync and polling scripts for automated workflows, full API reference
Retrieve code review results from DeepSource using the DeepSource CLI — issues, vulnerabilities, report cards, and analysis runs.
- Use when: you want to list code review issues, bugs, or security findings for a repository
- Use when: you want to check dependency vulnerabilities or CVE status
- Use when: you want to get code review grades, coverage metrics, or report cards
- Use when: you want to view analysis run history
Install all skills:
npx skills add deepsource/skills
Install a specific skill:
npx skills add deepsource/skills -s autofix-bot-api
npx skills add deepsource/skills -s deepsource
Skills work with 18+ AI agents including Claude Code, GitHub Copilot, Cursor, Cline, Codex, and more.
Once installed, skills are automatically available to your AI agent. Just ask naturally:
- "Scan this repo for security vulnerabilities using Autofix Bot"
- "Run Autofix Bot on my latest changes and fix any issues"
- "Upload this repository to Autofix Bot and analyze it"
- "Show me the critical code issues in this repo"
- "Check for dependency vulnerabilities"
- "What's the report card for this repository?"
Each skill follows the Agent Skills specification:
skill-name/
├── SKILL.md # Instructions and metadata (required)
├── scripts/ # Executable helper scripts (optional)
├── references/ # Detailed reference documentation (optional)
└── assets/ # Static resources (optional)
MIT