The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the uv tool to execute a local Python script (scripts/mcpcall.py) whenever a user requests library resolution or documentation fetching.
[EXTERNAL_DOWNLOADS]: The helper script scripts/mcpcall.py defines standard Python dependencies (mcp, anyio, httpx, httpx-sse) in its metadata block, which are downloaded from official package registries at runtime.
[DATA_EXFILTRATION]: The skill connects to https://mcp.context7.com/mcp to retrieve documentation. This process transmits the CONTEXT7_API_KEY provided in the environment variable to the service provider for authentication via a Bearer token.
[PROMPT_INJECTION]: The skill acts as a surface for indirect prompt injection because it ingests third-party documentation and code examples from an external API.
Ingestion points: Results from session.call_tool in scripts/mcpcall.py are output directly to the agent's context.
Boundary markers: No explicit delimiters or instructions are used to warn the agent about potential instructions embedded in the external documentation.
Capability inventory: The script performs network requests and prints output; the risk involves the agent potentially following malicious instructions contained within the fetched content.
Sanitization: Content from the API is processed and printed without filtering or escaping.

context7