The Agent Skills Directory

[DATA_EXFILTRATION]: The skill scans the filesystem for sensitive file paths, including private keys (.pem, .key), environment files (.env, .env.*), and database files (.sqlite, .db). Although this is done to ensure these files are ignored by version control, it identifies the locations of potential credentials within the repository context.
[COMMAND_EXECUTION]: The skill uses git and fd to list files and directories. It explicitly searches for hidden and sensitive files using the -H flag to provide context for the ignore patterns.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by processing filenames and repository structures which are untrusted external data.
Ingestion points: Filesystem scanning via git ls-files and fd as defined in SKILL.md.
Boundary markers: None; the skill does not use delimiters or instructions to isolate file paths from the agent's logic processing.
Capability inventory: The skill can execute shell commands and modify the .gitignore file.
Sanitization: No validation or sanitization is performed on the gathered file paths before they are used in the agent's reasoning or output.

gitignore