What we do
Unit tests catch code regressions.
Spooled catches behavioral regressions.
01
Behavioral Fingerprinting
Every agent run produces a structural fingerprint from its execution shape — tool sequence, call graph, error patterns. Three-dimensional similarity scoring catches drift that output evals miss.
02
Privacy-Safe Architecture
Spooled captures execution structure — call sequences, timing, tool names — not the content of your calls. Prompts, responses, and tool payloads are stripped at the SDK level before storage.
03
Policy-Gated Merges
YAML rules that block PRs on behavioral regressions. Enforce tool allowlists, set latency budgets, require baselines. Your guardrails, version-controlled alongside your code.
© 2026 GMH Labs LLC. All rights reserved.