churn-reduction

Pass

Audited by Gen Agent Trust Hub on Sep 25, 2026

Risk Level: SAFEPROMPT_INJECTIONCREDENTIALS_UNSAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill implements a robust defense-in-depth strategy by providing explicit instructions to ignore and report any prompt injection attempts found in fetched or pasted content.
  • [CREDENTIALS_UNSAFE]: Explicit prohibitions are in place against asking for, persisting, or echoing credentials. The skill mandates redaction of tokens and session identifiers from all outputs.
  • [INDIRECT_PROMPT_INJECTION]: The skill has a data ingestion surface through fetching external websites and processing user-provided data. This is addressed by comprehensive security guidelines in the reference files that require quoting third-party content and using HTML escaping to prevent cross-site scripting (XSS).
  • [COMMAND_EXECUTION]: No unauthorized command execution patterns were detected. The skill uses standard platform tools for information retrieval and visualization.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 25, 2026, 04:06 PM