<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0"><channel><title>Decision Horizon blog</title><link>https://decisionhorizon.dev/blog/</link><description>Notes on agent evidence.</description><language>en</language><item><title>Decision Horizon for Claude Code, Cursor, Copilot and every coding agent</title><link>https://decisionhorizon.dev/blog/coding-agents/</link><guid>https://decisionhorizon.dev/blog/coding-agents/</guid><pubDate>Sun, 04 Oct 2026 18:00:00 +0000</pubDate><description>Coding agents act through built-in tools that no MCP proxy sees. One command now puts a policy decision, an Ostiary score and a signed record on every one of their tool calls.</description></item><item><title>Ostiary: a doorkeeper for every agent action</title><link>https://decisionhorizon.dev/blog/introducing-ostiary/</link><guid>https://decisionhorizon.dev/blog/introducing-ostiary/</guid><pubDate>Sun, 04 Oct 2026 16:00:00 +0000</pubDate><description>Rules catch what you thought to write down. Ostiary learns how each agent normally behaves, scores every tool call before it runs, and signs its reasons into the record.</description></item><item><title>Authorization decides. Evidence proves.</title><link>https://decisionhorizon.dev/blog/authorization-decides-evidence-proves/</link><guid>https://decisionhorizon.dev/blog/authorization-decides-evidence-proves/</guid><pubDate>Sun, 04 Oct 2026 11:00:00 +0000</pubDate><description>The agent-security field has converged on checking every tool call before it runs. That stops the bad action. Proving later why an action was allowed takes something more.</description></item><item><title>Capture at decision time</title><link>https://decisionhorizon.dev/blog/capture-at-decision-time/</link><guid>https://decisionhorizon.dev/blog/capture-at-decision-time/</guid><pubDate>Sun, 04 Oct 2026 10:00:00 +0000</pubDate><description>Decision provenance, bitemporal data, and why reconstructing an agent&#x27;s decision after the fact goes wrong. The most trustworthy evidence is the record written at the moment.</description></item><item><title>The hosted record store and the Console</title><link>https://decisionhorizon.dev/blog/hosted-record-store-and-console/</link><guid>https://decisionhorizon.dev/blog/hosted-record-store-and-console/</guid><pubDate>Sun, 04 Oct 2026 09:00:00 +0000</pubDate><description>Your records can now live in a hosted store you never have to trust for the truth, and your security team can search and verify them in a browser.</description></item><item><title>What it would take to make the record lie</title><link>https://decisionhorizon.dev/blog/what-it-would-take-to-make-the-record-lie/</link><guid>https://decisionhorizon.dev/blog/what-it-would-take-to-make-the-record-lie/</guid><pubDate>Sat, 03 Oct 2026 09:00:00 +0000</pubDate><description>Tamper-evident is not tamper-proof. Here is what an attacker would need for each kind of lie, and what every record says about itself.</description></item><item><title>What an Agent Decision Record contains</title><link>https://decisionhorizon.dev/blog/what-an-agent-decision-record-contains/</link><guid>https://decisionhorizon.dev/blog/what-an-agent-decision-record-contains/</guid><pubDate>Fri, 02 Oct 2026 09:00:00 +0000</pubDate><description>Eight blocks and one signature: who sponsored the action, what authority it held, what the agent had read, which rule let it through, and what the record cannot see.</description></item><item><title>Introducing Decision Horizon</title><link>https://decisionhorizon.dev/blog/introducing-decision-horizon/</link><guid>https://decisionhorizon.dev/blog/introducing-decision-horizon/</guid><pubDate>Thu, 01 Oct 2026 09:00:00 +0000</pubDate><description>When an AI agent does something consequential, nobody can say why it was allowed. We built a signed record of what it knew, was allowed, and did, at the moment it acted.</description></item></channel></rss>
