The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes bash commands to record usage metrics, including timestamps and the current git repository name, to a local log file located at ~/.gstack/analytics/skill-usage.jsonl.
[COMMAND_EXECUTION]: The skill uses eval to process the output of a local binary script (~/.claude/skills/gstack/bin/gstack-paths) to dynamically configure environment variables for the session.
[PROMPT_INJECTION]: The skill accepts a directory path from the user and uses it within a shell command to set an edit boundary, creating an indirect prompt injection surface.
Ingestion points: User-provided directory path requested via the AskUserQuestion tool in SKILL.md.
Boundary markers: None identified.
Capability inventory: The skill utilizes Bash, Read, Write, and Edit tools to enforce boundaries.
Sanitization: The input is placed inside double quotes in the shell resolution step (cd "<user-provided-path>"), which provides basic protection but does not explicitly sanitize against shell command substitution characters like backticks or dollar signs.

guard