The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes a local bash script (vet.sh) used to scan directories for potentially dangerous patterns using the grep utility. This execution is limited to scanning files and reporting findings.
[INDIRECT_PROMPT_INJECTION]: The skill's primary function is to process and analyze untrusted source code from other skills, which creates a potential surface for indirect prompt injection.
Ingestion points: The agent is instructed in SKILL.md to read the full content of SKILL.md and any referenced scripts in the target skill directory.
Boundary markers: The instructions do not include specific delimiters or warnings to treat the reviewed code as data rather than instructions, though the context of 'vetting' implies a critical review.
Capability inventory: The agent uses the vet.sh script to perform searches and must generate a verdict based on its analysis of the code. The agent also performs manual inspection of code logic as instructed in SKILL.md.
Sanitization: There is no evidence of sanitization or filtering of the external content before it is processed by the agent.

skill-vetting