incident-commander

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data, such as incident descriptions and event logs, and interpolates this content directly into generated reports and summaries. This creates a surface for indirect prompt injection where an attacker could embed malicious instructions in log files to influence the behavior of an AI agent that processes the resulting reports.
  • Ingestion points: The scripts scripts/incident_classifier.py, scripts/pir_generator.py, and scripts/timeline_reconstructor.py ingest user-supplied data via stdin or file inputs.
  • Boundary markers: The templates for generating documents do not use delimiters or isolation markers to distinguish user-provided data from system instructions.
  • Capability inventory: The Python scripts use only the standard library and lack dangerous capabilities such as network access or shell execution.
  • Sanitization: The scripts lack specific sanitization for prompt injection patterns when interpolating input into templates.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 04:56 PM
Security Audit — agent-trust-hub — incident-commander