The Agent Skills Directory

[SAFE]: The skill implements standard development workflows for code maintenance and cleanup. The functionality is well-defined and scoped to local repository management.
[INDIRECT_PROMPT_INJECTION]: The skill processes local source files, which presents a surface for indirect prompt injection if external or untrusted code is being analyzed.
Ingestion points: Reads local files using Read, Grep, and Glob tools.
Boundary markers: None identified in the prompt instructions to distinguish code from instructions.
Capability inventory: Includes Edit for file modification and Bash for command execution.
Sanitization: No specific sanitization or filtering of the ingested code content is performed before processing.

techdebt