The Agent Skills Directory

[COMMAND_EXECUTION]: The skill invokes multiple vendor-provided utilities from the ~/.claude/skills/gstack/bin/ directory, including gstack-skill-start, gstack-slug, gstack-learnings-search, gstack-paths, and gstack-learnings-log. It uses the eval command to integrate the output of gstack-slug and gstack-paths into the execution environment.
[DYNAMIC_EXECUTION]: In the implementation of the stats command, the skill uses bun -e to execute a hardcoded JavaScript snippet that processes and aggregates data from the learnings.jsonl file.
[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection through the ingestion of historical project data.
Ingestion points: Project learning data is read from the learnings.jsonl file via the cat command and the gstack-learnings-search utility.
Boundary markers: The skill does not implement explicit delimiters or instructions to ignore potential commands embedded within the retrieved learning insights.
Capability inventory: The skill has access to the Bash tool for executing shell commands and the Write and Edit tools for modifying files in the repository.
Sanitization: There is no evidence of sanitization or validation performed on the insight field or other data retrieved from the project history before it is processed or presented.

learn