Score Agent Response Quality
PassAudited by ClawScan on May 6, 2026.
Overview
This is a coherent instruction-only response-scoring skill, with the main caution that pasted agent responses may contain sensitive information.
This skill appears safe to use for evaluating agent replies. Before installing or invoking it, redact PII, credentials, customer details, and confidential business data from any response you paste. Also treat the monetization-readiness output as rubric-based guidance, not as a guarantee of revenue or platform acceptance.
Findings (1)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
If users paste PII, credentials, customer data, or confidential business content, that text may enter the agent conversation context even though the skill itself shows no external data access.
The skill may handle sensitive pasted text and includes a local-processing privacy assurance. Because no external tooling or code is shown and the skill advises redaction, this is a user-awareness note rather than a concern.
If the user pastes a response that contains user PII, suggest they redact before pasting. The skill processes everything locally, but good hygiene is good hygiene.
Redact sensitive or regulated information before using the skill, and treat the local-processing statement as limited to the provided instruction-only artifacts.