The Agent Skills Directory

[COMMAND_EXECUTION]: The skill invokes shell commands including git status, git diff, and pnpx difit to identify code changes and launch the visualization tool.
[EXTERNAL_DOWNLOADS]: The skill uses pnpx to execute the difit package, which involves fetching the package from a public registry at runtime. This behavior is documented as functional since it targets the primary tool defined for this skill.
[DYNAMIC_EXECUTION]: The skill generates JSON-formatted comments at runtime and passes them as arguments to the difit command-line tool.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the local repository that could contain malicious instructions designed to influence the AI's explanation or the generated comments.
Ingestion points: The skill reads output from git status, git diff, and file contents from the local file system (SKILL.md, Step 1 and 2).
Boundary markers: The skill does not define specific delimiters or instructions for the AI to ignore embedded commands within the diffs it processes.
Capability inventory: The skill has the ability to execute shell commands via pnpx and access the local file system via git.
Sanitization: There is no evidence of sanitization or filtering of the content retrieved from the code diffs before it is processed by the AI.

difit-explain