The Agent Skills Directory

[COMMAND_EXECUTION]: The skill interpolates the $ARGUMENTS variable directly into a Bash shell command in SKILL.md without proper quoting. This pattern is susceptible to shell injection, where a user can execute arbitrary commands by including shell metacharacters (e.g., ;, &&, |) in the input arguments.
Evidence: Found in SKILL.md within the uv run and fallback python3 command lines: uv run "${CLAUDE_SKILL_DIR}/scripts/okf_validate.py" $ARGUMENTS.
[EXTERNAL_DOWNLOADS]: The skill's fallback execution path installs the pyyaml library from the Python Package Index (PyPI) at runtime.
Evidence: python3 -m pip install --quiet pyyaml in SKILL.md.
[PROMPT_INJECTION]: The skill processes untrusted Markdown content from a user-specified directory, creating a surface for indirect prompt injection where instructions embedded in the data could influence the agent's behavior.
Ingestion points: scripts/okf_validate.py recursively reads and parses all .md files within the provided bundle directory.
Boundary markers: No markers or explicit instructions are provided to the agent to ignore instructions embedded within the bundle files.
Capability inventory: The skill has access to the Bash tool, which is used to execute the Python validator and manage dependencies.
Sanitization: No sanitization or filtering is performed on the Markdown content or frontmatter before the script processes it and outputs findings.

validate