eval-core-scorecard
Pass
Audited by Gen Agent Trust Hub on Feb 19, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [NO_CODE] (SAFE): The skill consists entirely of markdown-based logic and configuration in SKILL.md and agents/openai.yaml. No Python, Node.js, or shell scripts are present, eliminating the risk of direct malicious code execution.
- [SAFE] (SAFE): No indicators of obfuscation, credential exposure, or persistence mechanisms were found. The skill operates as a structural wrapper for other evaluation tools.
- [Indirect Prompt Injection] (SAFE): The skill possesses a surface for indirect prompt injection as it ingests untrusted assistant responses, but it is not exploitable due to a lack of system-level capabilities. 1. Ingestion points: assistant response text and conversation context in SKILL.md. 2. Boundary markers: Absent; the workflow does not define delimiters for user input. 3. Capability inventory: None; the skill only generates JSON output and does not access the network, file system, or shell. 4. Sanitization: Absent.
Audit Metadata