Check a tool result for schema drift, empty success, and silent truncation before the agent acts.…
Check a tool result for schema drift, empty success, and silent truncation before the agent acts. An OpenClaw agent is about to write, send, or hand off on a tool payload. Run this first. Catches schema drift, empty success (ok true but empty data), and silent truncation. Command: python3 {baseDir}/scripts/audit_tool_result.py RESULT.json --schema SCHEMA.json [--out REPORT.md] Exit 0 clean, 1 findings, 2 bad input. Stdlib only. No network. Scanner 84.