spec-test-suites

Pass

Audited by Gen Agent Trust Hub on Mar 6, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: No attempts to override system prompts or bypass safety guidelines were detected in the instructions or metadata.
  • [DATA_EXFILTRATION]: No network operations, such as curl or fetch, were found. The skill does not contain hardcoded credentials and only accesses local project documentation files.
  • [REMOTE_CODE_EXECUTION]: The skill does not download or execute any external scripts or packages.
  • [COMMAND_EXECUTION]: There are no shell commands, subprocess calls, or scripts that would execute code on the host system.
  • [INDIRECT_PROMPT_INJECTION]: While the skill reads untrusted data from project files (e.g., usecase.md), it lacks dangerous capabilities like network access or code execution that could be exploited via injection.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 6, 2026, 08:45 AM