skills/noih/dev-skills/delegate/Gen Agent Trust Hub

delegate

Pass

Audited by Gen Agent Trust Hub on Nov 18, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill defines a protocol for processing reports and findings from delegated subagents, which introduces an indirect prompt injection surface where malicious instructions could be embedded in the data returned by these external processes.
  • Ingestion points: Delegated agent reports, diffs, evidence locations, and findings results as described in SKILL.md.
  • Boundary markers: The skill explicitly instructs the agent to verify claims by reading source evidence rather than forwarding reports, and to accept work against original requirements.
  • Capability inventory: The driver agent retains full access to file manipulation, command execution (via delegation tools), and subagent orchestration.
  • Sanitization: The instructions require manual verification of evidence and comparison with the original task record to prevent unverified data from influencing agent decisions.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 18, 2026, 05:35 PM