The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill defines a mandatory workflow where the agent must research external documentation using tools like Exa (via fuse-ai-pilot:research-expert) and mcp__context7__query-docs. This creates a vulnerability where malicious instructions embedded in external content could influence the agent's actions.
Ingestion points: External research content retrieved from the web via search and documentation query tools as specified in the Agent Workflow section of SKILL.md.
Boundary markers: The instructions do not provide delimiters or instructions to treat external content as untrusted data.
Capability inventory: The skill instructs the agent to execute shell commands (npx astro add) and write to project configuration files (astro.config.ts, tsconfig.json).
Sanitization: There are no verification or sanitization steps defined for the data retrieved during the research phase.
[COMMAND_EXECUTION]: The skill uses the npx astro add command to automate the addition of frameworks like React, Vue, and Svelte. This involves executing a CLI tool that modifies the local file system and installs dependencies.
[EXTERNAL_DOWNLOADS]: The skill installs several official and community-maintained packages from npm, including @astrojs/react, @astrojs/vue, @astrojs/svelte, @astrojs/solid-js, @astrojs/preact, @astrojs/alpinejs, @astrojs/lit, @qwikdev/astro, and @analogjs/astro-angular. These are well-known resources within the Astro ecosystem.
[REMOTE_CODE_EXECUTION]: The use of npx to install and run the Astro CLI integration scripts involves downloading and executing code from the npm registry. As these are official or well-known community packages for a reputable framework, they are considered standard developer dependencies.

astro-integrations