The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes Node.js scripts to automate query classification, documentation fetching, and result analysis. It also instructs the agent to use shell commands such as cat for file piping, git clone for repository access, and npm install -g for tool setup.
[EXTERNAL_DOWNLOADS]: The skill fetches documentation content from context7.com and library-specific websites. It also supports cloning external GitHub repositories for deep code analysis when structured documentation is unavailable.
[INDIRECT_PROMPT_INJECTION]: The skill processes technical documentation (llms.txt) retrieved from external, third-party websites, creating a surface for indirect prompt injection attacks.
Ingestion points: Documentation is fetched via scripts/fetch-docs.js from context7.com and other external library domains.
Boundary markers: No explicit boundary markers or instructions to disregard embedded prompts are provided when the agent processes the external content.
Capability inventory: The agent has the ability to execute Node.js scripts, perform shell operations, and install global NPM packages.
Sanitization: While URLs are parsed by the scripts, the textual content of the documentation is not sanitized or filtered for potential malicious instructions before analysis by the agent.

docs-seeker