The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The mcp.json file configures the agent to download the gitnexus package from the npm registry at runtime using npx.
[REMOTE_CODE_EXECUTION]: The skill executes external code by running npx -y gitnexus@latest mcp, which fetches and runs the most recent version of the package without a fixed version pin.
[COMMAND_EXECUTION]: The gitnexus_rename tool provides functionality for automated modification of multiple files in the local filesystem, allowing the agent to perform broad changes to the project.
[COMMAND_EXECUTION]: Instructions in SKILL.md suggest that the user run npx gitnexus analyze in their terminal, which executes an external binary on the host system.
[PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection as it processes codebase content (untrusted data) and uses it to generate file modification plans.
Ingestion points: Codebase analysis tools like gitnexus_impact, gitnexus_query, and gitnexus_context read source code from the local environment into the agent context.
Boundary markers: No explicit markers or instructions to ignore embedded commands in the processed code are present.
Capability inventory: The skill possesses file-write capabilities through gitnexus_rename and provides terminal execution instructions.
Sanitization: There is no evidence of sanitization or validation of the content read from the codebase before it influences the agent's actions.

gitnexus-refactoring