verification

Fail

Audited by Gen Agent Trust Hub on Jul 20, 2026

Risk Level: HIGHCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill provides instructions to execute PowerShell scripts (scripts/check_skill_integrity.ps1 and scripts/research_quality_gate.ps1) using the -ExecutionPolicy Bypass argument. This specifically instructs the agent to bypass Windows PowerShell execution policies, which are security controls intended to prevent the execution of untrusted scripts.
  • [PROMPT_INJECTION]: The skill processes untrusted data from external sources during its verification workflow, creating a surface for indirect prompt injection.
  • Ingestion points: The skill reads file content from research projects, chapters, and external API responses such as the CrossRef API.
  • Boundary markers: The instructions lack specific delimiters or system instructions to ignore potential commands embedded within the external data being verified.
  • Capability inventory: The agent is granted the ability to read files, use grep, and execute PowerShell scripts with bypassed security policies.
  • Sanitization: No sanitization or validation logic is specified to filter or escape the external content before it is processed by the agent.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Jul 20, 2026, 01:58 AM
Security Audit — agent-trust-hub — verification