Learn
Each article is a 3-6 minute read with concrete code, no sales fluff. Pick one based on what you're trying to evaluate.
-
How the SBO3L playground works (3-tier architecture)
6 minFor: Anyone trying to evaluate SBO3L from the playground
Why the playground splits into mock cinematic, WASM client-side, and hosted live daemon — three tiers for three audiences.
-
How the audit chain prevents tampering
4 minFor: Engineers evaluating cryptographic audit guarantees
SHA-256 + JCS canonical JSON + append-only — why a single byte-flip breaks the whole chain.
-
On-chain anchoring: closing the local-tamper gap
4 minFor: Auditors + compliance teams + sponsor reviewers
Hash-chained logs detect local tampers. On-chain anchors detect wholesale chain replacement. 24K gas per anchor; 6h cron.
-
MEV guard — three layers of slippage defense
3 minFor: Treasury automation + DEX-trading agents
Slippage cap + recipient allowlist + private mempool requirement. What gets denied, what doesn't.
-
Why LangChain needs SBO3L (and what changes when you wire them)
5 minFor: LangChain devs already shipping agents in production
LangChain wraps your agent's reasoning. SBO3L wraps LangChain's tool calls in policy + signed audit. Five lines of code, one boundary your CFO understands.
-
Trust DNS Manifesto — naming as authentication for autonomous agents
22 minFor: ENS standards reviewers, ERC-8004 implementers, agent-platform architects
RFC-style 5000-word manifesto. Why SBO3L resolves ENS names to trust commitments rather than wallet addresses; the seven-record opinionated profile; the standardisation path through ENSIP-N.
-
Don't give your agent a wallet. Don't make KeeperHub guess what's authorized either.
8 minFor: Agent-platform engineers + KeeperHub workflow authors evaluating end-to-end safety
Composing SBO3L's policy boundary with KeeperHub's execution layer. Why two products beat one monolith, the signed receipt as the contract, IP-1..IP-5, end-to-end demo, the 15 issues we filed, joint roadmap.