Decision-Theory

Make Your OpenClaw Cheaper and Harder to Fool

credence-pi is an OpenClaw plugin plus a local daemon that learns your agent's behaviour and governs its tool calls: exact-repeat waste blocked at precision and recall 1.0, an injected exfiltration surfaced as a confirmation at 0.94 precision. Research-stage, local-first, installable today.

essaysbayesianaidecision-theory

What a Regex Can't Do

The governor from 'The Brain is Opaque to the Body' now ships as an OpenClaw plugin: wasted tool calls blocked at precision and recall 1.0 on real sessions, injected exfiltration surfaced as a confirmation at 0.94 precision. And an argument: several of these behaviours are beyond any hand-tuned heuristic, because matching them re-derives Bayesian decision theory.

essaysbayesianaidecision-theory

The Brain is Opaque to the Body

A first-pass body-brain architecture for governing a coding agent's tool_call hook with a Bayesian decision-theoretic brain. The wire schema is fixed by what Pass 1 ships; Pass 2 swaps the posterior representation without disturbing it. Plus what credence-lint and a precedent system caught at the eleventh hour.

essaysbayesianaidecision-theory