resilience-audit

Pass

Audited by Gen Agent Trust Hub on Apr 16, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it is designed to ingest and process untrusted application source code and live web content.\n
  • Ingestion points: Source code files are read during the exploration phase (Phase 2), and live web content is accessed during interactive verification (Phase 4).\n
  • Boundary markers: Prompts provided for sub-agents (e.g., in references/agent-prompts.md) do not utilize explicit delimiters or safety instructions to distinguish between the skill's instructions and the untrusted content being analyzed.\n
  • Capability inventory: The skill maintains the ability to write to the local filesystem (reports/resilience-audit.md) and execute browser-based tasks via platform-specific tools (tabs_context_mcp, tabs_create_mcp).\n
  • Sanitization: No validation or sanitization mechanisms are implemented for the data retrieved from the audited codebase or web pages before it is processed or reported.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 16, 2026, 01:41 AM