signals-scout-session-replay
Pass
Audited by Gen Agent Trust Hub on Oct 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted data originating from end-user browsers, such as URLs, element text, and console logs, which could contain malicious instructions intended to manipulate the agent.
- Ingestion points: Data enters the context via the
eventstable (specifically$current_url,$el_text, and console messages) and through the outputs of replay vision scanners. - Boundary markers: The instructions explicitly direct the agent to quote untrusted snippets and treat session content strictly as data rather than instructions.
- Capability inventory: The skill utilizes
execute-sqlfor data analysis andemit_report/edit_reportfor surfacing findings. - Sanitization: The skill mandates aggressive truncation and the use of sanitized identifiers (slugified paths or element labels) when storing data in the scratchpad or generating reports to mitigate potential injection vectors.
Audit Metadata