<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0"><channel>
<title>Theo Martin</title><link>https://theomart.in</link><atom:link xmlns:atom="http://www.w3.org/2005/Atom" href="https://theomart.in/feed.xml" rel="self"/><description>Writing on coding agents</description><language>en</language>
<item><title>What Claude Code training costs in France</title><link>https://theomart.in/en/writing/claude-code-training-in-france/</link><guid>https://theomart.in/en/writing/claude-code-training-in-france/</guid><pubDate>Tue, 25 Aug 2026 00:00:00 +0000</pubDate><description>Public prices from five French training vendors, collected in late August 2026, and what a crawl of their catalogs actually proves.</description></item>
<item><title>Every agent works in the demo</title><link>https://theomart.in/en/writing/every-agent-works-in-the-demo/</link><guid>https://theomart.in/en/writing/every-agent-works-in-the-demo/</guid><pubDate>Sat, 28 Mar 2026 00:00:00 +0000</pubDate><description>The resumption contract: what a fresh session reads first, why conversation history is not state, and the test that separates a demo from production.</description></item>
<item><title>Most SKILL.md files work once</title><link>https://theomart.in/en/writing/most-skill-md-files-work-once/</link><guid>https://theomart.in/en/writing/most-skill-md-files-work-once/</guid><pubDate>Sat, 28 Mar 2026 00:00:00 +0000</pubDate><description>What makes an agent skill hold up past the first session: a trigger-clear description, a short file, scripts for deterministic work, a resumption contract.</description></item>
<item><title>The 99% problem</title><link>https://theomart.in/en/writing/the-99-percent-problem/</link><guid>https://theomart.in/en/writing/the-99-percent-problem/</guid><pubDate>Sat, 28 Mar 2026 00:00:00 +0000</pubDate><description>Compound error rates in long agent pipelines, the success table at 99, 95 and 90 percent per step, and why shorter segments beat better models.</description></item>
<item><title>The model is the easy part</title><link>https://theomart.in/en/writing/the-model-is-the-easy-part/</link><guid>https://theomart.in/en/writing/the-model-is-the-easy-part/</guid><pubDate>Sat, 28 Mar 2026 00:00:00 +0000</pubDate><description>Scaffolding moves task completion more than model choice does, and what a production agent harness has to hold up under.</description></item>
<item><title>Your agent isn&#x27;t confused. It&#x27;s drowning.</title><link>https://theomart.in/en/writing/your-agent-isnt-confused-its-drowning/</link><guid>https://theomart.in/en/writing/your-agent-isnt-confused-its-drowning/</guid><pubDate>Sat, 28 Mar 2026 00:00:00 +0000</pubDate><description>Context bloat as the main undiagnosed failure mode: tool overload, raw corpora handed to agents, and subagents used as context firewalls.</description></item>
<item><title>The context envelope: what your LLM chat agent needs but isn&#x27;t getting</title><link>https://theomart.in/en/writing/the-context-envelope/</link><guid>https://theomart.in/en/writing/the-context-envelope/</guid><pubDate>Wed, 25 Mar 2026 00:00:00 +0000</pubDate><description>The JSON payload a web app should wrap around every chat turn: six sections, the prompt injection and cache gotchas, and a ship order to follow.</description></item>
<item><title>The MCP paradox</title><link>https://theomart.in/en/writing/the-mcp-paradox/</link><guid>https://theomart.in/en/writing/the-mcp-paradox/</guid><pubDate>Wed, 25 Mar 2026 00:00:00 +0000</pubDate><description>A year of MCP: the integration problem it solved, the context bloat it created, the CLI comeback, and why the real constraint on agents is attention.</description></item>
<item><title>TypeScript vs JavaScript for AI agents: the feedback loop is everything</title><link>https://theomart.in/en/writing/typescript-vs-javascript-for-ai-agents/</link><guid>https://theomart.in/en/writing/typescript-vs-javascript-for-ai-agents/</guid><pubDate>Wed, 25 Mar 2026 00:00:00 +0000</pubDate><description>The compiler gives a coding agent a loop it can close on its own. Without it you are the loop, and the regression and token numbers show the gap.</description></item>
<item><title>Your new job title is AI office manager</title><link>https://theomart.in/en/writing/your-new-job-title-is-ai-office-manager/</link><guid>https://theomart.in/en/writing/your-new-job-title-is-ai-office-manager/</guid><pubDate>Wed, 25 Mar 2026 00:00:00 +0000</pubDate><description>Most of the day around a coding agent is plumbing. What compounds is the CLAUDE.md files, the skills, and the connectors that make a workspace usable.</description></item>
</channel></rss>
