The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a local bash script located at ${CLAUDE_PLUGIN_ROOT}/scripts/index-query.sh to query documentation status. This is a functional requirement for the skill's stated purpose.
[PROMPT_INJECTION]: The skill ingests the current directory name via $(pwd) to use as a search parameter. While this provides a surface for indirect prompt injection via malicious directory names, the risk is mitigated by the skill's restricted scope of querying a local index.
Ingestion points: The working directory path enters the agent context via the pwd command (SKILL.md).
Boundary markers: The search term is enclosed in double quotes within the shell command, though no explicit instructions are provided to ignore embedded commands in the path string.
Capability inventory: The skill can execute local bash scripts via the shell (SKILL.md).
Sanitization: The skill relies on the basename utility but does not perform additional sanitization on the path string before execution.

project-entry