The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to perform security reviews on external codebases, which constitutes a surface for indirect prompt injection where malicious instructions embedded in the analyzed code could attempt to influence the agent's behavior.
Ingestion points: Accesses external files and data flows using Read, Grep, and Glob tools as specified in the workflow.
Boundary markers: The skill does not define specific delimiters or instructions to treat the analyzed file content as untrusted data or to ignore embedded instructions.
Capability inventory: The skill is permitted to use the Bash tool for command execution, which could be targeted by an injection attack.
Sanitization: There are no documented mechanisms for sanitizing or escaping the content read from files before it is processed by the agent.
[SAFE]: The skill does not contain any hardcoded credentials, obfuscated code, or unauthorized network operations.
[SAFE]: The instructions explicitly guide the agent to perform read-only operations and follow security best practices, such as identifying hardcoded secrets and validating inputs.

security-review