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