skills/utkusen/sast-skills/sast-rce/Gen Agent Trust Hub

sast-rce

Pass

Audited by Gen Agent Trust Hub on Sep 20, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted source code, creating an attack surface where malicious comments or data within the codebase could attempt to influence the agent's behavior.
  • Ingestion points: Codebase files and architecture documentation (sast/architecture.md) are processed as context for analysis.
  • Boundary markers: The skill implements a robust three-phase pipeline (recon, verification, and merge) that requires specific subagent instructions and standardized markdown templates, creating structural boundaries for the analysis.
  • Capability inventory: The skill performs filesystem write operations to store analysis results in the sast/ directory.
  • Sanitization: The skill provides extensive documentation on 'What RCE is NOT' and 'Secure Examples' for multiple programming languages, which serves to semantically guide the agent and minimize the chance of misinterpreting malicious data as instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 20, 2026, 03:34 AM
Security Audit — agent-trust-hub — sast-rce