terraform-consumer-design

Pass

Audited by Gen Agent Trust Hub on Mar 3, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: Executes a local shell script at .specify/scripts/bash/validate-env.sh to verify credentials and environment state before initiating the design process.\n- [COMMAND_EXECUTION]: Utilizes the GitHub CLI (gh) to create and modify issues, where parameters are derived from user-provided natural language input.\n- [PROMPT_INJECTION]: Ingests untrusted natural language requests from users to populate structured GitHub issue templates and guide subagent tasks, creating a surface for indirect prompt injection.\n
  • Ingestion points: User infrastructure requests are captured via the $ARGUMENTS variable and mapped to multiple template fields (e.g., infrastructure_components, project_name).\n
  • Boundary markers: The instructions do not define delimiters or specific 'ignore embedded instructions' warnings when processing user input.\n
  • Capability inventory: The agent can execute local bash scripts, perform Git operations, use the gh CLI, and invoke various subagents (speckit.specify, speckit.plan, etc.) to generate technical artifacts.\n
  • Sanitization: There is no evidence of sanitization, escaping, or validation of user-provided strings before they are incorporated into commands or subagent prompts.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 3, 2026, 02:17 AM