Observed arrival · 2026-09-21
Pum Doctrine Wants You to Audit the AI
Credibility concern recorded. The source reference remains available for verification and correction.
A browser- and terminal-based verification layer for AI actions, claims, limits, receipts, and system records.
- For
- AI builders auditing autonomous systems
- Worth noticing
- It exposes a tamper sandbox alongside Ed25519, SHA-256, Merkle inclusion, and cross-machine root checks.
Field notes
The project organizes its evidence around signed receipts, chained records, and a manifest whose files can be checked locally. Its browser workflow uses built-in cryptography, while the terminal path supplies OpenSSL, curl, and Python commands. The page also describes a fifteen-hash Merkle proof and a second machine that witnesses the record root. These are concrete verification mechanisms, although the extract cannot independently confirm the broader claims about the AI system or the live state measurements.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
One card from the complete issue