careful

Pass

Audited by Gen Agent Trust Hub on Aug 7, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill defines a PreToolUse hook that executes bin/check-careful.sh. This script is used to validate bash commands before they are run by the agent.
  • [DATA_EXFILTRATION]: The skill maintains a local usage log at ~/.gstack/analytics/skill-usage.jsonl. It records the repository name and timestamps. No network calls are present to exfiltrate this data; it remains on the local machine.
  • [SAFE]: Heuristic detections of destructive commands within the script are false positives. The strings rm -rf and kubectl delete are used as regex patterns to protect the user, not to perform unauthorized deletions.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 7, 2026, 08:36 PM
Security Audit — agent-trust-hub — careful