The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill's primary function is to ingest and analyze external SKILL.md files provided by the user. It lacks boundary markers or explicit system-level instructions to ignore potential commands embedded within the analyzed files.
Ingestion points: The skill reads the content of SKILL.md and directory structures from arbitrary local paths provided in the user's prompt (Steps 1, 2, and 4 in SKILL.md).
Boundary markers: Absent. There are no instructions to delimit the untrusted content or warn the agent against following instructions found within the data.
Capability inventory: The skill possesses the ability to execute shell commands (ls, sed via extract-yaml.sh), run Python scripts (skill-audit.py), and write new markdown files (reports) to the local filesystem.
Sanitization: Absent. Content from external files is parsed and evaluated directly as text.
[COMMAND_EXECUTION]: The skill relies on local scripts to perform its analysis, which involve command execution.
Evidence: SKILL.md instructs the agent to use scripts/extract-yaml.sh (which uses sed) and scripts/skill-audit.py (which uses the re and pathlib Python libraries) to process local files. It also uses ls -la to verify directory structures.

skill-quality-reviewer