principle-explain-the-number
Pass
Audited by Gen Agent Trust Hub on Oct 12, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill consists entirely of natural language guidelines designed to improve the agent's reasoning when interpreting performance metrics and evaluation results.- [NO_CODE]: There are no scripts, binaries, or shell commands associated with this skill. It operates solely as a prompt-based principle.- [INDIRECT_PROMPT_INJECTION]: While the skill processes external data (measurement results), it lacks any operational capabilities (network, file-write, or code execution) that could be exploited through malicious data injection.- [METADATA_POISONING]: The YAML frontmatter is used correctly to define the skill's name and purpose, with no evidence of deceptive instructions in the metadata fields.
Audit Metadata