openspec-apply-change

Pass

Audited by Gen Agent Trust Hub on Jun 24, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes the openspec CLI tool with various arguments (list, status, instructions) to retrieve workflow data in JSON format. This is the intended behavior of the skill for managing implementation tasks.
  • [DATA_EXPOSURE]: The skill reads project files identified by the CLI as contextFiles. This access is necessary for the agent to understand the codebase and requirements before implementing changes.
  • [COMMAND_EXECUTION]: The skill instructs the agent to make code changes and update task lists (e.g., marking checkboxes in markdown files). These operations are scoped to the implementation of the specified tasks.
  • [INDIRECT_PROMPT_INJECTION]: Because the skill processes external data (files and CLI output) to determine its actions, it has a theoretical exposure to indirect prompt injection. However, it includes guardrails such as requiring clarification for ambiguous tasks and keeping changes minimal and focused.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 24, 2026, 06:38 AM
Security Audit — agent-trust-hub — openspec-apply-change