chaos-experiment

Pass

Audited by Gen Agent Trust Hub on Mar 26, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it accepts and processes user-supplied data for infrastructure identifiers and experiment definitions.
  • Ingestion points: The org_id, project_id, and body parameters in harness_list, harness_create, and harness_execute tools within SKILL.md.
  • Boundary markers: The instructions do not define clear delimiters or use 'ignore embedded instructions' warnings for the data being interpolated into the MCP tool calls.
  • Capability inventory: The skill leverages tools capable of modifying and deleting cloud infrastructure resources (e.g., K8s pods, EC2 instances) as part of its core functionality.
  • Sanitization: No sanitization or validation logic is specified for the body content or identifier strings before they are utilized by the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 26, 2026, 10:22 PM
Security Audit — agent-trust-hub — chaos-experiment