iterative-retrieval
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill implements a pattern where the agent retrieves and evaluates files from a repository, which creates an attack surface for indirect prompt injection. Malicious instructions hidden in source code files could manipulate the agent during context evaluation or subsequent task execution.
- Ingestion points: The
retrieveFilesfunction and the evaluation loop described inSKILL.mdingest arbitrary codebase content. - Boundary markers: The pattern does not include specific delimiters or instructions to ignore embedded commands within retrieved files.
- Capability inventory: The skill assumes the agent has file-reading capabilities and can perform follow-up actions like bug fixing or feature implementation.
- Sanitization: There is no requirement or logic provided for sanitizing or escaping the content retrieved from external files.
Audit Metadata