The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides instructions for the agent to execute shell commands to inspect local project files and run development utilities like the Mastra CLI.
Evidence: Shell commands such as ls node_modules/@mastra/, cat node_modules/@mastra/core/dist/docs/SOURCE_MAP.json, and npm run dev are used to verify local installation state and project configuration.
[EXTERNAL_DOWNLOADS]: The skill retrieves framework documentation and guides from official remote sources.
Evidence: Fetches content from https://mastra.ai/llms.txt and various subpages under https://mastra.ai/reference/ to ensure current API usage.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process information from external documentation and local package files, which presents a theoretical injection surface.
Ingestion points: Local node_modules documentation files and remote mastra.ai documentation URLs.
Boundary markers: None explicitly defined in the fetch or read commands.
Capability inventory: Shell command execution (cat, ls, grep, npm).
Sanitization: No explicit sanitization or filtering of documentation content is described.
[REMOTE_CODE_EXECUTION]: The skill recommends using npx to execute official framework migration and setup tools.
Evidence: Mentions npx @mastra/codemod@latest v1 in the version migration guide.

mastra