Skills
skills/bacoco/bmad-skills/openspec-change-proposal/Gen Agent Trust Hub

openspec-change-proposal

Fail

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: HIGHPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • PROMPT_INJECTION (HIGH): The skill is susceptible to Indirect Prompt Injection due to how it processes external data. Ingestion points: Untrusted user input via change_request and impact_surface. Boundary markers: None; there are no instructions or delimiters to isolate untrusted content from the agent's logic. Capability inventory: The skill has access to Bash, Write, Read, and Grep tools. Sanitization: No sanitization or validation of the input content is performed before processing.
  • COMMAND_EXECUTION (MEDIUM): The skill uses the Bash tool to execute local scripts. Evidence: The process section explicitly invokes scripts/scaffold_change.py. While the script is intended to be local, execution of scripts with potentially user-influenced parameters (<change-id>) increases the attack surface for command injection.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Feb 16, 2026, 08:59 PM