The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes standard Unix utilities including grep, cat, tail, head, sed, and diff to manipulate text files. These operations are restricted to the repository context and are used to implement the core logic of upgrading documentation files.
[INDIRECT_PROMPT_INJECTION]: The skill processes content from CODING_STANDARDS.md files located in target repositories. While these files are external inputs, the risk of indirect prompt injection is mitigated by the mechanical nature of the shell-based processing and the disable-model-invocation: true configuration, which prevents the agent from interpreting the content as instructions. Furthermore, the skill enforces a human-in-the-loop requirement by showing a diff and waiting for approval before writing changes.
[DATA_EXPOSURE]: The skill accesses and reads project-level markdown files. It does not attempt to access sensitive system paths, hidden environment files, or credentials.

my-skills-standards-update