validating-completed-tasks

Pass

Audited by Gen Agent Trust Hub on Nov 12, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill implements robust security guardrails, including instructions to verify runtime environments (e.g., image digests and mounts), avoid unauthorized destructive operations (e.g., volume resets), and maintain a clear boundary between theoretical risks and demonstrated exploits.
  • [INDIRECT_PROMPT_INJECTION]: The skill identifies a potential attack surface where untrusted data (task descriptions and colleague reports) could contain malicious instructions. 1. Ingestion points: SKILL.md and research-deliverables.md. 2. Boundary markers: Instructions to distinguish mandatory behavior from author examples and identify real critical paths. 3. Capability inventory: File reads and delegated execution of test flows. 4. Sanitization: Mandatory redaction of credentials and secrets in SKILL.md (Step 5).
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 12, 2026, 07:03 PM