The Agent Skills Directory

[SAFE]: The skill's instructions focus on legitimate software engineering practices like refactoring, type safety, and technical debt reduction.
[COMMAND_EXECUTION]: The skill utilizes common development toolchains (e.g., pnpm, knip, madge) for codebase analysis and verification. These operations are conducted within the context of the user's project and do not exhibit malicious patterns.
[EXTERNAL_DOWNLOADS]: The skill explicitly states that missing tools should only be installed with user approval, which mitigates the risk of unauthorized external downloads.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it processes repository content to drive its logic.
Ingestion points: The skill reads and processes source code, comments, and configuration files from the repository being analyzed.
Boundary markers: No specific delimiters or safety warnings are used when the agent context ingests repository data.
Capability inventory: The skill has the capability to modify files and execute terminal commands through the package manager (pnpm).
Sanitization: The skill does not perform sanitization of ingested code before it is processed by the agent lanes.

deslop