The Agent Skills Directory

[SAFE]: No malicious patterns detected. The skill provides architectural and research guidance for AI agents.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest data from external sources such as npm, PyPI, and GitHub search results. While this is an attack surface for indirect prompt injection, it is the primary intended purpose of a research skill and is considered low risk in this context.
[COMMAND_EXECUTION]: The skill suggests using standard tools like rg, ls, and package managers (npm, pip) to inspect the local environment and project dependencies. These are used for information gathering to support the research workflow.
[DATA_EXPOSURE_AND_EXFILTRATION]: The skill mentions checking ~/.claude/settings.json and local skill directories. This is intended to help the agent identify available tools and MCP servers, which is a legitimate part of the agent's operation and does not involve exfiltrating sensitive data.

search-first