The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to process external content such as proposals, paths, or plans provided via the $ARGUMENTS variable.
Ingestion points: Untrusted data enters the agent context through the $ARGUMENTS parameter in SKILL.md and through the review pages it is instructed to reconcile.
Boundary markers: There are no explicit instructions to the agent to ignore or delimit embedded instructions within the ingested content.
Capability inventory: The skill executes shell-based logic including the running of check-playbooks.mjs and complex playbook instructions.
Sanitization: No sanitization or validation of the input content is specified before it is processed by the LLM.
[COMMAND_EXECUTION]: The skill instructs the agent to execute a local JavaScript file as part of its workflow.
Evidence: The instructions explicitly require the agent to run check-playbooks.mjs as defined in the repository's documentation.

verdict