mutant-survival-triage

Pass

Audited by Gen Agent Trust Hub on Sep 4, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external data from mutation testing reports, which constitutes an indirect prompt injection attack surface.
  • Ingestion points: The skill reads Stryker JSON, PIT XML, mutmut output, and Mull JSON reports to extract details like mutated code snippets and mutator descriptions (as defined in SKILL.md and references/tool-normalization.md).
  • Boundary markers: No explicit boundary markers or instructions to ignore embedded commands are present in the prompts to distinguish between report data and agent instructions.
  • Capability inventory: The skill is strictly read-only and does not possess capabilities for file writing, network exfiltration, or shell command execution. Its primary output is the generation of test proposals for human review.
  • Sanitization: There is no evidence of input validation or sanitization for the content processed from these external reports.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 4, 2026, 07:07 AM
Security Audit — agent-trust-hub — mutant-survival-triage