iterative-retrieval

Pass

Audited by Gen Agent Trust Hub on Apr 16, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • [SAFE]: The skill consists entirely of instructional markdown defining a logic flow for information retrieval.
  • [NO_CODE]: There are no executable scripts, binaries, or configuration files that could perform unauthorized actions.
  • [INDIRECT_PROMPT_INJECTION]: The protocol involves processing external data (transcripts, documents, codebases). While this is a known attack surface, the skill includes a maximum loop limit (Phase 4) to prevent infinite cycles.
  • Ingestion points: External sources such as transcripts, documents, and codebases mentioned in Phase 1 of SKILL.md.
  • Boundary markers: Absent; the protocol does not specify the use of delimiters when processing retrieved content.
  • Capability inventory: Subagent dispatch and resumption (resuming via agent ID).
  • Sanitization: Absent; the protocol does not detail specific sanitization steps for the retrieved information.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 16, 2026, 08:51 AM