cli-review-runner

Pass

Audited by Gen Agent Trust Hub on Apr 12, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes a target binary or script provided by the user to probe its behavior (e.g., testing help output and error responses). This execution is performed with safety wrappers, including a 5-second wall-clock timeout and stdin redirection from /dev/null, minimizing the risk of hangs or unintended interactivity.
  • [SAFE]: No indicators of malicious behavior, prompt injection, data exfiltration, or obfuscation were found. The skill also manages an indirect prompt injection surface safely: 1. Ingestion points: CLI stdout and stderr are read in probes.sh; 2. Boundary markers: Data is encapsulated in NDJSON findings; 3. Capability inventory: Process execution is restricted to the target CLI; 4. Sanitization: Characters are escaped via the crr_json_escape function in common.sh.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 12, 2026, 10:17 PM