observe
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a local JavaScript tool (
observe.mjs) using Node.js. This script is part of the tool's framework and is used to process local hook events and session data. - [INDIRECT_PROMPT_INJECTION]: The skill processes and reports on data from local session logs (hook events). This represents a potential injection surface if the logs contain untrusted content; however, the instructions are limited to reporting rule compliance and summary statistics, which limits the risk.
- Ingestion points: Local session logs and hook events read by the
observe.mjsscript. - Boundary markers: None explicitly defined in the instructions for the resulting output.
- Capability inventory: Executes Node.js scripts via shell commands; performs file existence checks using
ls. - Sanitization: The instructions do not specify sanitization for the log output, but the agent is directed to quote specific summary lines.
Audit Metadata