spec

Pass

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: LOW
Full Analysis
  • PROMPT_INJECTION (SAFE): The skill contains standard instructional guidelines for a development workflow. There are no attempts to override system prompts, bypass safety filters, or use 'Ignore previous instructions' patterns.
  • DATA_EXFILTRATION (SAFE): No network operations (curl, wget, fetch) or access to sensitive file paths (secrets, SSH keys) were detected. The file paths mentioned are relative project paths (e.g., locales, core standards).
  • EXTERNAL_DOWNLOADS (SAFE): No external package installations or script downloads are present. While it mentions tools like OpenSpec and Spec Kit, it does not provide commands to download or install them from untrusted sources.
  • REMOTE_CODE_EXECUTION (SAFE): There are no patterns involving the execution of remote scripts or piping web content to interpreters.
  • COMMAND_EXECUTION (SAFE): The skill suggests workflow commands (e.g., /openspec proposal) which appear to be internal tool aliases or user-facing documentation rather than automated shell execution. No dangerous OS commands (sudo, rm -rf) are present.
  • OBFUSCATION (SAFE): No Base64, zero-width characters, homoglyphs, or encoded strings were found.
  • DATA_EXPOSURE (SAFE): No hardcoded credentials, API keys, or private tokens are present in the documentation or templates.
Audit Metadata
Risk Level
LOW
Analyzed
Feb 16, 2026, 08:20 AM