Skip to content

an open-source alternative to the dots, bots and muses of the world, run an office of claude code/codex like agents on your laptop, sandboxes or anywhere, uses your existing subscriptions

seo-engine: stable sitemap lastmod, blog section in llms.txt, plan mode links

chaitanyagiri956c032dbd3f
diff --git a/blog/eleventy.config.js b/blog/eleventy.config.js
index a2d3c3c1..eedc5810 100644
--- a/blog/eleventy.config.js
+++ b/blog/eleventy.config.js
@@ -134,6 +134,14 @@ export default function (eleventyConfig) {
eleventyConfig.addFilter("htmlDate", (d) =>
DateTime.fromJSDate(d, { zone: "utc" }).toFormat("yyyy-LL-dd")
);
+ // Newest publish or update date across posts. The sitemap uses it for the
+ // home page lastmod so a rebuild on a later day does not change the file.
+ eleventyConfig.addFilter("newestPostDate", (posts) =>
+ (posts || []).reduce((max, p) => {
+ const d = new Date(p.data.updated || p.date);
+ return d > max ? d : max;
+ }, new Date(0))
+ );
// Reading time from rendered HTML / raw content (~225 wpm).
eleventyConfig.addFilter("readingTime", (content) => {
diff --git a/blog/scripts/postbuild.mjs b/blog/scripts/postbuild.mjs
index 036dfb9f..c9441761 100644
--- a/blog/scripts/postbuild.mjs
+++ b/blog/scripts/postbuild.mjs
@@ -6,7 +6,7 @@
// uses absolute URLs, so its location on disk doesn't affect its contents.
//
// Runs with cwd = blog/ (npm sets it to the package dir).
-import { rename } from "node:fs/promises";
+import { readFile, rename, rm, writeFile } from "node:fs/promises";
const from = "../docs/blog/sitemap.xml";
const to = "../docs/sitemap.xml";
@@ -18,3 +18,25 @@ try {
console.error(`[postbuild] failed to move sitemap: ${err.message}`);
process.exit(1);
}
+
+// Splice the generated blog section (src/llms-blog.njk) into the root llms.txt
+// between the BLOG markers, so AI crawlers see every post without anyone
+// editing llms.txt by hand. The temporary file is removed after.
+const section = "../docs/blog/llms-blog-section.txt";
+const llms = "../docs/llms.txt";
+const START = "<!-- BLOG:START generated by blog/scripts/postbuild.mjs -->";
+const END = "<!-- BLOG:END -->";
+try {
+ const body = (await readFile(section, "utf8")).trim();
+ const text = await readFile(llms, "utf8");
+ const a = text.indexOf(START);
+ const b = text.indexOf(END);
+ if (a === -1 || b === -1 || b < a) throw new Error("BLOG markers missing in docs/llms.txt");
+ const next = text.slice(0, a + START.length) + "\n" + body + "\n" + text.slice(b);
+ if (next !== text) await writeFile(llms, next);
+ await rm(section);
+ console.log(`[postbuild] blog section → docs/llms.txt`);
+} catch (err) {
+ console.error(`[postbuild] failed to update llms.txt: ${err.message}`);
+ process.exit(1);
+}
diff --git a/blog/src/llms-blog.njk b/blog/src/llms-blog.njk
new file mode 100644
index 00000000..0a9abb4b
--- /dev/null
+++ b/blog/src/llms-blog.njk
@@ -0,0 +1,16 @@
+---
+permalink: /llms-blog-section.txt
+eleventyExcludeFromCollections: true
+---
+{#- Blog section for the root llms.txt. scripts/postbuild.mjs splices this
+ between the BLOG markers in docs/llms.txt and deletes this file. -#}
+## Blog
+
+{{ collections.posts.length }} posts, grouped by topic, newest first in each topic. Topic hubs: {{ site.origin }}{{ '/topics/' | u }}
+{% for c in collections.categories %}
+### {{ c.name }}
+
+{% for p in c.posts -%}
+- [{{ p.data.title | striptags | trim | safe }}]({{ site.origin }}{{ p.url | u }}): {{ p.data.description | striptags | replace("\n", " ") | trim | safe }}
+{% endfor -%}
+{% endfor %}
\ No newline at end of file
diff --git a/blog/src/posts/aider-vs-claude-code.md b/blog/src/posts/aider-vs-claude-code.md
index 6a98e2d8..a321903b 100644
--- a/blog/src/posts/aider-vs-claude-code.md
+++ b/blog/src/posts/aider-vs-claude-code.md
@@ -33,7 +33,7 @@ In Aider you add the files you want changed to the chat, and it sends the model
{% img "note-1" %}
-Claude Code also has subagents with their own context window, tools and permissions, hooks around every tool call, MCP servers, and a plan mode that researches without editing. Aider's extras are more manual: an architect mode where one model plans and another writes the edits, `/run` and `/test` to feed command output back, and `--watch-files` for AI coding comments left in your files.
+Claude Code also has subagents with their own context window, tools and permissions, hooks around every tool call, MCP servers, and a [plan mode](/blog/how-to-use-claude-code-plan-mode/) that researches without editing. Aider's extras are more manual: an architect mode where one model plans and another writes the edits, `/run` and `/test` to feed command output back, and `--watch-files` for AI coding comments left in your files.
## What does a fresh Aider install show?
diff --git a/blog/src/posts/claude-code-max-plan-tips.md b/blog/src/posts/claude-code-max-plan-tips.md
index e4a82357..b588d8cd 100644
--- a/blog/src/posts/claude-code-max-plan-tips.md
+++ b/blog/src/posts/claude-code-max-plan-tips.md
@@ -54,7 +54,7 @@ Keep the conversation short, because Claude Code resends it with every request.
* Keep `CLAUDE.md` under 200 lines and move workflow instructions into skills, which load only when used.
* Run `/context` to see what is loaded, and disable MCP servers you are not using.
-Plan before editing. Shift+Tab into plan mode for anything that touches several files, press Escape the moment Claude heads the wrong way, and use `/rewind` instead of stacking "try again" on a bad attempt. Each retry rereads everything before it.
+Plan before editing. Shift+Tab into [plan mode](/blog/how-to-use-claude-code-plan-mode/) for anything that touches several files, press Escape the moment Claude heads the wrong way, and use `/rewind` instead of stacking "try again" on a bad attempt. Each retry rereads everything before it.
## Does prompt caching help on a subscription?
diff --git a/blog/src/posts/what-is-vibe-coding.md b/blog/src/posts/what-is-vibe-coding.md
index db27e440..bb971ce4 100644
--- a/blog/src/posts/what-is-vibe-coding.md
+++ b/blog/src/posts/what-is-vibe-coding.md
@@ -53,7 +53,7 @@ It bites in three places.
Keep the speed of prompting and add a check wherever a mistake gets expensive. Three habits cover most of it.
-**Plan before editing.** Ask the agent for a plan and approve it before any file changes. In Claude Code that is plan mode, where it reads files and proposes a plan but makes no edits until you approve ([Claude Code common workflows](https://code.claude.com/docs/en/common-workflows)). We tried it on 14 Sep 2026 with Claude Code 2.1.270 in an empty folder (output trimmed):
+**Plan before editing.** Ask the agent for a plan and approve it before any file changes. In Claude Code that is [plan mode](/blog/how-to-use-claude-code-plan-mode/), where it reads files and proposes a plan but makes no edits until you approve ([Claude Code common workflows](https://code.claude.com/docs/en/common-workflows)). We tried it on 14 Sep 2026 with Claude Code 2.1.270 in an empty folder (output trimmed):
```
$ claude -p --permission-mode plan --model sonnet "Create a file called hello.txt containing the word hi."
diff --git a/blog/src/sitemap.njk b/blog/src/sitemap.njk
index 984dd725..435dd0ff 100644
--- a/blog/src/sitemap.njk
+++ b/blog/src/sitemap.njk
@@ -14,7 +14,7 @@ eleventyExcludeFromCollections: true
<urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9">
<url>
<loc>{{ site.origin }}/</loc>
- <lastmod>{{ build.date }}</lastmod>
+ <lastmod>{{ collections.posts | newestPostDate | htmlDate }}</lastmod>
<changefreq>weekly</changefreq>
<priority>1.0</priority>
</url>
diff --git a/docs/blog/aider-vs-claude-code/index.html b/docs/blog/aider-vs-claude-code/index.html
index 092a4526..8467751c 100644
--- a/docs/blog/aider-vs-claude-code/index.html
+++ b/docs/blog/aider-vs-claude-code/index.html
@@ -194,7 +194,7 @@
<p>Aider is a pair programmer you steer file by file, and Claude Code is an agent that goes looking for itself.</p>
<p>In Aider you add the files you want changed to the chat, and it sends the model a repo map: a ranked outline of classes and function signatures across the repo, 1k tokens by default. The model replies in an edit format Aider picks per model, usually search and replace blocks, and Aider applies the edit and commits it to git, so <code>/undo</code> takes it back. Claude Code starts from the task. It searches and reads files with its own tools, edits them, runs commands such as your tests, and commits when you ask. Instead of a commit per edit, it snapshots files before each prompt so <code>/rewind</code> can roll a turn back.</p>
<figure class="fig fig-inline"><img src="/blog/assets/media/aider-vs-claude-code/note-1.png" alt="Hand-drawn sketch from “Aider vs Claude Code in 2026: Which Should You Use?”" loading="lazy" decoding="async" /></figure>
-<p>Claude Code also has subagents with their own context window, tools and permissions, hooks around every tool call, MCP servers, and a plan mode that researches without editing. Aider’s extras are more manual: an architect mode where one model plans and another writes the edits, <code>/run</code> and <code>/test</code> to feed command output back, and <code>--watch-files</code> for AI coding comments left in your files.</p>
+<p>Claude Code also has subagents with their own context window, tools and permissions, hooks around every tool call, MCP servers, and a <a href="/blog/how-to-use-claude-code-plan-mode/">plan mode</a> that researches without editing. Aider’s extras are more manual: an architect mode where one model plans and another writes the edits, <code>/run</code> and <code>/test</code> to feed command output back, and <code>--watch-files</code> for AI coding comments left in your files.</p>
<h2 id="what-does-a-fresh-aider-install-show" tabindex="-1">What does a fresh Aider install show? <a class="anchor" href="#what-does-a-fresh-aider-install-show" aria-hidden="true">#</a></h2>
<p>It shows version 0.86.2 and two git defaults worth knowing. We installed Aider into a throwaway folder with <code>uv tool install --python 3.12 aider-chat</code>, which took 28 seconds, then ran only version and help commands, with no key and no prompt:</p>
<pre><code>$ aider --version
diff --git a/docs/blog/claude-code-max-plan-tips/index.html b/docs/blog/claude-code-max-plan-tips/index.html
index 05f92013..d6e32204 100644
--- a/docs/blog/claude-code-max-plan-tips/index.html
+++ b/docs/blog/claude-code-max-plan-tips/index.html
@@ -206,7 +206,7 @@
<li>Keep <code>CLAUDE.md</code> under 200 lines and move workflow instructions into skills, which load only when used.</li>
<li>Run <code>/context</code> to see what is loaded, and disable MCP servers you are not using.</li>
</ul>
-<p>Plan before editing. Shift+Tab into plan mode for anything that touches several files, press Escape the moment Claude heads the wrong way, and use <code>/rewind</code> instead of stacking “try again” on a bad attempt. Each retry rereads everything before it.</p>
+<p>Plan before editing. Shift+Tab into <a href="/blog/how-to-use-claude-code-plan-mode/">plan mode</a> for anything that touches several files, press Escape the moment Claude heads the wrong way, and use <code>/rewind</code> instead of stacking “try again” on a bad attempt. Each retry rereads everything before it.</p>
<h2 id="does-prompt-caching-help-on-a-subscription" tabindex="-1">Does prompt caching help on a subscription? <a class="anchor" href="#does-prompt-caching-help-on-a-subscription" aria-hidden="true">#</a></h2>
<p>Yes, while the cache is warm. Claude Code’s prompt caching page says the main conversation gets a one hour cache on a subscription within plan usage, dropping to five minutes once you draw on usage credits, and subagents get five minutes. We checked by reading the <code>usage</code> fields Claude Code 2.1.270 writes to its session transcripts on our machine, with no cache settings changed. Across 27 requests in an orchestrating session, all 121,041 cache write tokens were one hour writes; a subagent it started wrote only five minute entries over 69 requests, and both read 96% to 98% of input from cache on 14 Sep 2026. To check yours, run <code>claude -p &quot;hello&quot; --output-format json</code> and read <code>usage.cache_creation</code>.</p>
<p>So come back inside the hour, or <code>/clear</code> and start fresh, rather than reheating a huge stale context. Switching model mid task rebuilds the cache, and on most models so does changing effort, so choose both at the start. Our guide to <a href="/blog/prompt-caching-for-ai-agents/">prompt caching for AI agents</a> explains why a stable prefix matters.</p>
diff --git a/docs/blog/what-is-vibe-coding/index.html b/docs/blog/what-is-vibe-coding/index.html
index aad3202f..18297276 100644
--- a/docs/blog/what-is-vibe-coding/index.html
+++ b/docs/blog/what-is-vibe-coding/index.html
@@ -206,7 +206,7 @@
</ul>
<h2 id="how-do-you-vibe-code-without-the-damage" tabindex="-1">How do you vibe code without the damage? <a class="anchor" href="#how-do-you-vibe-code-without-the-damage" aria-hidden="true">#</a></h2>
<p>Keep the speed of prompting and add a check wherever a mistake gets expensive. Three habits cover most of it.</p>
-<p><strong>Plan before editing.</strong> Ask the agent for a plan and approve it before any file changes. In Claude Code that is plan mode, where it reads files and proposes a plan but makes no edits until you approve (<a href="https://code.claude.com/docs/en/common-workflows">Claude Code common workflows</a>). We tried it on 14 Sep 2026 with Claude Code 2.1.270 in an empty folder (output trimmed):</p>
+<p><strong>Plan before editing.</strong> Ask the agent for a plan and approve it before any file changes. In Claude Code that is <a href="/blog/how-to-use-claude-code-plan-mode/">plan mode</a>, where it reads files and proposes a plan but makes no edits until you approve (<a href="https://code.claude.com/docs/en/common-workflows">Claude Code common workflows</a>). We tried it on 14 Sep 2026 with Claude Code 2.1.270 in an empty folder (output trimmed):</p>
<pre><code>$ claude -p --permission-mode plan --model sonnet &quot;Create a file called hello.txt containing the word hi.&quot;
**Plan:** Create `hello.txt` in the working directory containing the text `hi`, then read it back to confirm.
$ ls
diff --git a/docs/llms.txt b/docs/llms.txt
index 334f489d..e2dc0e73 100644
--- a/docs/llms.txt
+++ b/docs/llms.txt
@@ -40,8 +40,199 @@
- [Claude Squad vs Munder Difflin](https://munderdiffl.in/blog/claude-squad-vs-munder-difflin/)
- [Best Claude Code multi-agent tools](https://munderdiffl.in/blog/best-claude-code-multi-agent-tools/)
+<!-- BLOG:START generated by blog/scripts/postbuild.mjs -->
+## Blog
+
+161 posts, grouped by topic, newest first in each topic. Topic hubs: https://munderdiffl.in/blog/topics/
+
+### guides
+
+- [How to Install and Use Munder Difflin: A Beginner's Guide](https://munderdiffl.in/blog/how-to-install-and-use-munder-difflin/): A plain language guide to Munder Difflin 0.5.2. What a coding agent is, which AI engine to pick based on what you already pay for, the one terminal command you'll ever run, and how to install and set it up on macOS, Windows or Linux.
+- [Your First Hour With Munder Difflin: Onboarding to a Pro Office](https://munderdiffl.in/blog/your-first-hour-with-munder-difflin/): Munder Difflin 0.5.2 from download to a working Pro office: install Claude Code, Codex or OpenCode, sign in, go Pro, then automate reviews, Slack and email.
+- [How to Get the Most Out of Your Claude Code Max Plan](https://munderdiffl.in/blog/claude-code-max-plan-tips/): Stretch a Claude Code Max plan: model choice, context hygiene, prompt caching, subagents and unattended runs, checked against Anthropic's own pages.
+- [How to Make the Most of Your Codex Max Plan](https://munderdiffl.in/blog/codex-max-plan-tips/): There is no Codex Max plan. What ChatGPT Pro 5x and 20x give you in Codex, how the weekly limit works, and how to stretch it.
+- [Build Your Personal AGI on Your Computer Using Claude Code and Munder Difflin](https://munderdiffl.in/blog/build-your-personal-agi-on-your-computer/): Turn Claude Code into a personal AGI on your own computer: agents with memory, schedules, Gmail, Calendar, voice and Slack, set up in Munder Difflin 0.5.2.
+- [How to Install Gemini CLI, and Who Can Still Use It in 2026](https://munderdiffl.in/blog/how-to-install-gemini-cli/): Install Google's Gemini CLI with npm on Mac, Windows or Linux, why Homebrew is stuck on 0.46.0, and which sign in still works after 18 Jun 2026.
+- [How to Connect Slack to Munder Difflin 0.5.2](https://munderdiffl.in/blog/connect-slack-to-munder-difflin/): Connect Slack to Munder Difflin 0.5.2: create the Slack app, add the bot scopes, pick one of three connection modes, and get answers in the thread.
+- [How to Connect Ollama to Claude Code](https://munderdiffl.in/blog/how-to-connect-ollama-to-claude-code/): Point Claude Code at Ollama with ollama launch claude or three environment variables, give the model 64K of context, and know what stops working.
+- [How to Install Codex CLI on Mac, Windows and Linux](https://munderdiffl.in/blog/how-to-install-codex-cli/): Install OpenAI's Codex CLI with the official script, npm or Homebrew, sign in with ChatGPT or an API key, then update it and check your version.
+- [Why Does Claude Code Keep Asking for Permission?](https://munderdiffl.in/blog/why-does-claude-code-keep-asking-for-permission/): Claude Code asks because Manual mode gates edits, shell commands and web requests. The fixes, safest first: modes, allow rules, sandboxing, then bypass.
+- [How to Use Claude Code Plan Mode](https://munderdiffl.in/blog/how-to-use-claude-code-plan-mode/): How Claude Code's plan mode works: turning it on, what it blocks, what a real headless plan mode run prints, and how to leave it.
+- [How to Add an MCP Server to Claude Code](https://munderdiffl.in/blog/how-to-add-an-mcp-server-to-claude-code/): The claude mcp add command, its scope flag, editing .mcp.json by hand, and how Munder Difflin's Capabilities screen grants an MCP server to one agent at a time.
+- [The Command Center: Kanban, Fleet and Budgets in One Place](https://munderdiffl.in/blog/command-center-guide/): A guide to Munder Difflin's Command Center in 0.5.2: the kanban with task dependencies, the ask me tab, triggers, memory, live token use against each agent's budget, and when to watch the board instead of the floor.
+- [How to Add a GitHub Copilot CLI Agent to Your Floor](https://munderdiffl.in/blog/how-to-add-a-github-copilot-cli-agent/): A step-by-step v0.3.3 walkthrough: install GitHub Copilot CLI (or let the harness install it for you), hire a Copilot worker, pick a model, and know exactly what print mode can and can't do.
+- [How to Hire From the Agent Gallery](https://munderdiffl.in/blog/how-to-hire-from-the-agent-gallery/): A practical guide to Munder Difflin's Agent Gallery: pick one of 80 ready made roles, download its manifest, import it with Add agent, review every field, customise identity, workspace, engine and briefing, then spawn it yourself.
+- [How to Brief Your Orchestrator (So the Floor Actually Ships)](https://munderdiffl.in/blog/how-to-talk-to-your-orchestrator/): A practical guide to briefing Munder Difflin's GOD orchestrator: state the goal not the steps, set constraints and budgets, name the deliverable, and let Michael staff the floor. With bad-vs-good brief examples, mid-run steering, and Talk mode.
+- [How to Use the Built-in Monaco IDE](https://munderdiffl.in/blog/how-to-use-the-built-in-monaco-ide/): A practical walkthrough of Munder Difflin v0.3.3's built-in Monaco IDE: the title-bar IDE button, the git CHANGES rail with side-by-side diffs vs HEAD, the file tree, tabs, Cmd/Ctrl+S save — and the agent review workflow it enables.
+- [Run a Mixed-Engine Office: the Right CLI for Every Desk](https://munderdiffl.in/blog/run-a-mixed-engine-office/): A practical guide to mixing agent engines on one Munder Difflin floor: Claude Code as orchestrator, Codex for coding bursts, Copilot for dispatched tasks, and OpenCode/Crush/pi.dev for BYOK keys and local models.
+- [Run Munder Difflin Locally on a Mac Mini](https://munderdiffl.in/blog/run-munder-difflin-on-a-mac-mini/): Run a whole Munder Difflin office offline on an Apple silicon Mac mini: size the model to your unified memory, install Ollama or LM Studio, and wire OpenCode, Crush, Qwen and Pi to it. Current as of Munder Difflin 0.5.2 and the M6 and M5 Pro Mac mini.
+- [Run Munder Difflin on Open Source Models: Fully Local or Through a Provider](https://munderdiffl.in/blog/run-munder-difflin-on-open-models/): Munder Difflin can run your whole agent floor on open weight models like gpt-oss, Qwen3, DeepSeek, Llama, GLM and Kimi: fully local with Ollama, LM Studio or vLLM, or through a provider with your own key. The wiring for each engine, current as of 0.5.2.
+- [Run Your AI Agent Hive from Slack: The Complete Setup](https://munderdiffl.in/blog/run-ai-agent-hive-from-slack-setup/): Step-by-step setup for driving Munder Difflin's AI agent hive from Slack: create the app, set scopes, paste the tunnel Request URL, and @mention your bot to start work — done-summaries post back in-thread.
+- [How to Trigger Your AI Agent Hive from a GitHub Webhook](https://munderdiffl.in/blog/trigger-ai-agents-from-github-webhook/): Wire a GitHub webhook to Munder Difflin: a secret-gated local endpoint turns each repo event into a task for your GOD orchestrator — POST a message, get a token, poll the result. No server to host.
+- [Munder Difflin v0.2.4 Feature Walkthrough](https://munderdiffl.in/blog/munder-difflin-v0-2-4-feature-walkthrough/): A comprehensive guide to every change in Munder Difflin v0.2.4 — how the Codex lifecycle-hook bridge achieves full hive parity, what the Schedules tab adds, why tunnelmole replaced localtunnel, and what else shipped.
+- [Does the June 2026 Agent SDK Change Affect Munder Difflin?](https://munderdiffl.in/blog/does-the-june-2026-agent-sdk-change-affect-munder-difflin/): From June 15, 2026 the Claude Agent SDK gets a separate credit. Munder Difflin drives the native Claude Code CLI, so your hive runs on your plan as before.
+- [MCP Tool Poisoning: How to Secure the Model Context Protocol](https://munderdiffl.in/blog/mcp-security-tool-poisoning/): How the Model Context Protocol gets attacked — tool poisoning, rug pulls, command injection — and the layered defenses that secure your MCP tools.
+- [The Multi-Agent Cost Playbook: Five Levers to Cut AI Agent Spend](https://munderdiffl.in/blog/the-multi-agent-cost-playbook/): Caching, batching, model tiering, context discipline, local-first — the five levers that cut a multi-agent fleet's bill, and how they compound.
+- [Running AI Agents Safely: Permission Modes and Sandboxing](https://munderdiffl.in/blog/agent-security-and-sandboxing/): How to run autonomous AI agents safely: permission modes and the bypassPermissions foot-gun, workspace isolation, and what to gate vs. allow.
+- [Building Reliable AI Agents: Failure Is the Default](https://munderdiffl.in/blog/building-reliable-ai-agents/): Agents fail constantly: tools error, steps stall, output goes wrong. Reliability isn't avoiding failure — it's making each one contained and recoverable.
+- [Claude Code Automation: Let Agents Build While You Sleep](https://munderdiffl.in/blog/claude-code-automation-while-you-sleep/): Claude Code automation that runs overnight: an autonomous Stop-hook loop, safe permission bypass, and guardrails so you wake up to progress, not chaos.
+- [Context Engineering for AI Agents: Budgeting the Window](https://munderdiffl.in/blog/context-engineering-for-ai-agents/): Why context engineering beats prompt wording for agents — and the tactics (isolation, retrieval, externalized state, compaction) that keep the window lean.
+- [How to Debug a Multi-Agent System](https://munderdiffl.in/blog/debugging-multi-agent-systems/): How to debug a multi-agent system: use the event log, per-agent terminals, message trails, and git history to find why a hive of AI agents went sideways.
+- [Evaluating AI Agent Reliability: How to Measure What You Can Trust](https://munderdiffl.in/blog/evaluating-ai-agent-reliability/): How to measure whether an AI agent is reliable enough to trust: reliability thresholds, why benchmarks overstate, and evaluating on your own codebase.
+- [Approving AI Agents: Human-in-the-Loop Without a Queue](https://munderdiffl.in/blog/human-in-the-loop-approving-ai-agents/): How to approve what AI agents do without a parallel approval queue: native permission prompts, routing 'to: human' to the orchestrator, remote approval.
+- [Prompt Caching for AI Agents: Stop Paying for the Same Tokens Twice](https://munderdiffl.in/blog/prompt-caching-for-ai-agents/): An agent re-sends the same prompt, tools, and context every turn. Prompt caching makes you pay for that prefix once — here's how to design for it.
+- [Security for AI Coding Agents: A Practical Guide](https://munderdiffl.in/blog/security-for-ai-coding-agents/): How to run AI coding agents safely: contain the blast radius, scope every task, treat agent input as untrusted, and gate the irreversible behind a human.
+- [A Small-Team Playbook for AI Coding Agents](https://munderdiffl.in/blog/small-team-ai-agents-playbook/): How a small team adopts a hive of AI coding agents without chaos — the roles to assign, the cadence to run, and the guardrails that keep it safe.
+- [The Lethal Trifecta for Coding Agents](https://munderdiffl.in/blog/the-lethal-trifecta-for-coding-agents/): A coding agent reads private code, ingests untrusted content, and runs commands — the lethal trifecta. How a poisoned dependency leaks your secrets.
+- [Trigger Your AI Agent Hive from Slack](https://munderdiffl.in/blog/trigger-ai-agents-from-slack/): Trigger an AI agent hive from Slack: a local webhook verifies each message and drops it into your orchestrator's queue as a task — no server to host.
+- [What's New in Agentic AI: A June 2026 Field Guide](https://munderdiffl.in/blog/whats-new-in-agentic-ai-june-2026/): A grounded June 2026 roundup of agentic AI: Claude Opus 4.8, the MCP and A2A protocol layer, usage-based agent billing, and agent governance.
+- [Bringing Your MCP Servers and Skills Into a Hive of Agents](https://munderdiffl.in/blog/mcp-and-skills-in-a-hive/): How agents in a hive inherit your existing Claude Code MCP servers, skills, and tools — and how to scope which agent gets which capability.
+- [How to Run Multiple Claude Code Agents (Without Losing Track)](https://munderdiffl.in/blog/how-to-run-multiple-claude-code-agents/): Run several Claude Code agents in parallel without the chaos: give each a role, let them coordinate, and stop alt-tabbing between terminal windows.
+- [How to Manage Multiple Claude Code Sessions Like a Pro](https://munderdiffl.in/blog/manage-multiple-claude-code-sessions/): Tactics to manage multiple Claude Code sessions without losing track: naming, roles, context isolation, and when a harness beats juggling terminal tabs.
+- [Set Up a Claude Code Multi-Agent Workflow in 10 Minutes](https://munderdiffl.in/blog/claude-code-multi-agent-setup-tutorial/): A from-zero Claude Code multi-agent setup: install Munder Difflin, onboard, spawn a few agents, and watch the GOD orchestrator route your first task.
+- [Claude Code Subagents vs a Multi-Agent Harness](https://munderdiffl.in/blog/claude-code-subagents-vs-multi-agent-harness/): Claude Code subagents vs a multi-agent harness: where subagents stop and a harness with shared memory, messaging, and an orchestrator takes over.
+
+### comparisons
+
+- [Grok Bot Alternatives: Six Picks, Grouped by What You Want](https://munderdiffl.in/blog/grok-bot-alternatives/): Grok Bot runs every Bot on one shared cloud computer. Six alternatives checked on 14 Sep 2026, grouped by hosting, privacy and model choice.
+- [Herdr Alternatives: What to Pick and Who Each One Fits](https://munderdiffl.in/blog/herdr-alternatives/): Herdr alternatives checked on 14 Sep 2026: tmux, Munder Difflin, Claude Squad, cmux, Orca, Conductor, Superset and Emdash, and who each fits.
+- [OmniAgent Alternatives: What to Use Instead of Databricks Omnigent](https://munderdiffl.in/blog/omniagent-alternatives/): OmniAgent is really Omnigent, the open source meta harness from Databricks. Alternatives checked on 14 Sep 2026, and who each one fits.
+- [Grok Bot Pricing: What $20, $30, $100 and $300 a Month Actually Buy](https://munderdiffl.in/blog/grok-bot-pricing/): Grok Bot has no plan of its own. It comes bundled with SuperGrok and Cursor plans from $20 to $300 a month. Here is what each one gets you, what the free trial is, and the $0 route for work agents.
+- [Meta Muse Alternatives: 7 Picks for People Who Can't (or Won't) Use Muse](https://munderdiffl.in/blog/meta-muse-alternatives/): Meta Muse is US only and trains on your chats by default. Seven alternatives, from open source agents that run on your own computer to hosted ones from OpenAI, Google and SpaceXAI.
+- [Muse and Grok Bot Want Your Passwords. Here Is Who Can See Them](https://munderdiffl.in/blog/meta-muse-grok-bot-privacy/): Meta Muse and Grok Bot both ask for your logins. We read both companies' security docs: who sees your passwords, what trains on your chats, and how to keep AI agents on your own computer instead.
+- [Meta Muse vs Grok Bot: Same Blueprint, Different Locks](https://munderdiffl.in/blog/meta-muse-vs-grok-bot/): Meta Muse and Grok Bot both give an AI agent its own cloud computer. We compared price, security, platforms and who each one is for, plus the free option that runs on your own machine.
+- [Top 10 Indian open source projects that are giving US big tech a tough time](https://munderdiffl.in/blog/indian-open-source-ai-projects-vs-big-tech/): Ten open source AI projects from Indian developers, each lined up against the big tech product it takes on. Laya vs Jev, Munder Difflin vs Muse and Grok Bot.
+- [Aider vs Claude Code in 2026: Which Should You Use?](https://munderdiffl.in/blog/aider-vs-claude-code/): Aider vs Claude Code, checked 15 Sep 2026: how each edits code, which models it runs, what it costs, MCP, and Aider's stalled releases.
+- [Claude Code vs Antigravity: Which One Fits Your Work?](https://munderdiffl.in/blog/claude-code-vs-antigravity/): Claude Code vs Google Antigravity, checked 14 Sep 2026: terminal agent or desktop agent app, Claude inside Antigravity, plans, limits and a verdict.
+- [Claude Code vs Cursor: Which Is Better in 2026?](https://munderdiffl.in/blog/claude-code-vs-cursor/): Claude Code vs Cursor, checked 14 Sep 2026: where each runs, models, plans, the new Claude Code weekly limits, parallel agents and using both.
+- [Codex CLI vs Claude Code](https://munderdiffl.in/blog/codex-cli-vs-claude-code/): Codex CLI and Claude Code compared firsthand: sandbox and approval flags, AGENTS.md vs CLAUDE.md, real pricing, and where each terminal agent wins.
+- [qm vs Munder Difflin: YC's Multiplayer Harness or an Office of Your Clones?](https://munderdiffl.in/blog/qm-vs-munder-difflin/): An honest September 2026 comparison of YC's qm and Munder Difflin. qm is a multiplayer agent harness a company deploys for Slack and the web. Munder Difflin runs an office of your clones on your own machine, with Teams to connect them.
+- [CrewAI and AutoGen vs a Local Agent Harness: Framework or App?](https://munderdiffl.in/blog/crewai-autogen-vs-a-local-agent-harness/): CrewAI, AutoGen (now Microsoft Agent Framework), and LangGraph are frameworks — Python you write to build your own agent system. A multi-agent harness is an app you download that runs a team on your repos today. Here's how to tell which one you actually need.
+- [GitHub Copilot CLI vs Claude Code as Hive Workers](https://munderdiffl.in/blog/github-copilot-cli-vs-claude-code-for-hive-work/): How GitHub Copilot CLI and Claude Code behave as agent engines inside a multi-agent harness: hooks, inbox mail, orchestration, session lifecycle — and why the right answer is to run both.
+- [Orca vs Munder Difflin: Agent IDE or Agent Office?](https://munderdiffl.in/blog/orca-vs-munder-difflin/): Orca is a YC-backed Agent IDE for driving coding agents side by side in isolated worktrees. Munder Difflin is an agent office that runs itself. An honest comparison of the two — and when each one is the right pick.
+- [Running Agents in tmux vs an Agent Harness](https://munderdiffl.in/blog/tmux-and-scripts-vs-an-agent-harness/): tmux panes, git worktrees, shell scripts, and cron will absolutely run several Claude Code sessions at once. Here's what that DIY setup does well, where it breaks at scale, and what a purpose-built agent harness automates.
+- [The Best AI Coding Agents in 2026: A Field Guide](https://munderdiffl.in/blog/best-ai-coding-agents/): A fair 2026 field guide to the best AI coding agents — Cursor, Aider, Cline, Devin, Copilot — by category, plus where a local multi-agent hive fits.
+- [Cline vs Munder Difflin: In-Editor Agent or Local Hive?](https://munderdiffl.in/blog/cline-vs-munder-difflin/): A fair comparison of Cline and Munder Difflin — an in-editor BYOK coding agent vs a local multi-agent orchestration hive — and when to pick which.
+- [The Best Tools to Run Multiple Claude Code Agents (2026)](https://munderdiffl.in/blog/best-claude-code-multi-agent-tools/): An honest September 2026 roundup of tools for running several Claude Code agents: Claude Code agent teams, Claude Squad, Conductor, Nimbalyst, Emdash, Vibe Kanban and Munder Difflin, and how to pick one.
+- [Claude Code Orchestration Tools, Compared](https://munderdiffl.in/blog/claude-code-orchestration-tools-compared/): Claude Code orchestration tools compared on memory, messaging, visibility, control, and local-first — across Claude Squad, Conductor, Crystal, and more.
+- [Claude Squad vs Munder Difflin: Which Should You Use?](https://munderdiffl.in/blog/claude-squad-vs-munder-difflin/): Claude Squad vs Munder Difflin: a lean terminal session manager against a memory-backed, orchestrated, visual hive. Feature table and an honest verdict.
+- [How AI Answer Engines Choose What to Cite](https://munderdiffl.in/blog/how-ai-answer-engines-choose-sources/): How ChatGPT, Claude, Perplexity, and Google AI Overviews actually pick and cite sources in 2026 — the per-engine signals, and why there's no single rank.
+- [Local-First vs Cloud Agent SDKs: Which Wins in 2026?](https://munderdiffl.in/blog/local-first-vs-cloud-agent-sdks/): Local-first agent hives vs the 2026 cloud agent SDK wave (OpenAI, Google ADK, Microsoft, MCP/A2A): what each optimizes for, and when to pick which.
+- [Looking for a Claude Squad Alternative?](https://munderdiffl.in/blog/claude-squad-alternative/): What Claude Squad does well, where it stops, and why a memory-backed, orchestrated hive may be the Claude Squad alternative you're after.
+- [A Conductor Alternative for Orchestrating Claude Code](https://munderdiffl.in/blog/conductor-claude-code-alternative/): Conductor is a polished macOS app for parallel Claude Code worktrees. When to pick a cross-platform, orchestrated hive as your Conductor alternative.
+- [A Crystal Alternative: Multi-Agent Claude Code with Memory](https://munderdiffl.in/blog/crystal-claude-code-alternative/): Crystal runs parallel Claude Code worktrees well. Where a memory-backed, orchestrated hive changes the workflow — and when to stick with Crystal.
+- [A vibe-kanban Alternative for Coordinating AI Agents](https://munderdiffl.in/blog/vibe-kanban-alternative/): vibe-kanban gives you a board to assign tasks to coding agents. When a self-routing hive is the better vibe-kanban alternative — and when a board wins.
+- [How to Choose a Multi-Agent Coding Tool: A Buyer's Checklist](https://munderdiffl.in/blog/how-to-choose-a-multi-agent-tool/): A decision framework for choosing a multi-agent coding tool — memory, control, visibility, cost, local-first — with a simple scoring rubric you can run.
+
+### concepts
+
+- [Is Ollama Good for Coding? What Local Models Can and Can't Do](https://munderdiffl.in/blog/is-ollama-good-for-coding/): Ollama only runs the model, so the answer depends on which one you pull and your memory. What local models do well, where Claude wins, and a verdict.
+- [Is It Worth Paying for Claude Code Max?](https://munderdiffl.in/blog/is-claude-code-max-worth-it/): Who should pay for Claude Max 5x or 20x, who should stay on Pro or an API key, and the break even math to run on your own Claude Code usage.
+- [Is Claude Code Open Source?](https://munderdiffl.in/blog/is-claude-code-open-source/): No. Claude Code is proprietary under Anthropic's terms. What its GitHub repo holds, what the 2026 leak changed, and which open tools sit around it.
+- [What Is Vibe Coding? Meaning, Risks and Safer Habits](https://munderdiffl.in/blog/what-is-vibe-coding/): Vibe coding means building software from AI prompts without reading the code. Where the term came from, when it is fine, where it bites, and what to check.
+- [What Is an MCP Server?](https://munderdiffl.in/blog/what-is-an-mcp-server/): An MCP server gives an AI agent tools, data and prompts over one open protocol. What it is, the host, client and server roles, and where Claude Code plugs in.
+- [How Much Does Claude Code Cost?](https://munderdiffl.in/blog/how-much-does-claude-code-cost/): What Claude Code actually costs: free plan access, Claude Pro and Max pricing, per token API rates, and what running several agents does to the bill.
+- [Loop Engineering: Designing Agent Loops That Converge](https://munderdiffl.in/blog/loop-engineering-for-ai-agents/): Everyone engineers the prompt; almost nobody engineers the loop around it. Stop conditions, drain loops, retry with backoff, compaction cycles, breaker escalation, budgets, and human gates — the outer-loop mechanics that make an agent converge instead of run away.
+- [Review Agent Work Where It Happens](https://munderdiffl.in/blog/review-agent-work-where-it-happens/): AI agents produce diffs faster than humans can review them, and alt-tabbing to an external editor breaks supervision flow. The case for review-in-place: Munder Difflin's built-in Monaco IDE puts a CHANGES rail and side-by-side diffs vs HEAD right on the office floor.
+- [Voice Is a Control Plane, Not a Gimmick](https://munderdiffl.in/blog/voice-as-a-control-plane-for-agent-fleets/): Voice is a terrible way to write code and a great way to run a fleet. Why low-bandwidth commands over high-bandwidth work is the right split — with Munder Difflin's Talk mode (echo-back confirmation, spend caps, michael-voice attribution) as the case study.
+- [What Is an Agent Development Environment (ADE)?](https://munderdiffl.in/blog/what-is-an-agent-development-environment/): ADE means two things in 2026: a platform for building agents (Letta's sense) and a workspace for shipping code with fleets of coding agents (Orca's sense). Here's the full tooling map — chat IDEs, agent CLIs, agent IDEs, and agent harnesses — and how to pick.
+- [What Is Harness Engineering?](https://munderdiffl.in/blog/what-is-harness-engineering/): Harness engineering is the discipline of building everything around the model — PTY plumbing, lifecycle hooks, mailboxes, memory, budgets, human gates, observability — because that's where agent reliability actually comes from. A definition, and four case studies from inside Munder Difflin.
+- [Shareable Agent Roles: A Portable Format for Hiring AI Coworkers](https://munderdiffl.in/blog/shareable-agent-roles/): Why agent roles should be portable: what a hire manifest encodes, how one-click hiring stays safe, and how The Hiring Fair turns tacit setup into a shareable artifact.
+- [The Clone Army Is a Trap: Why a Mixed Swarm Beats Ten Identical Agents](https://munderdiffl.in/blog/the-clone-army-trap-mixed-swarm-vs-identical-agents/): Spinning up tens of identical top-tier CLI agents feels powerful and burns tokens on work a cheaper agent could do. A mixed-capability swarm with shared memory wins on cost — and often on quality.
+- [Why CLI Agents Are So Powerful — and How Munder Difflin Spends Fewer Tokens Doing the Same Work](https://munderdiffl.in/blog/why-cli-agents-are-powerful/): CLI agents are powerful because they have terminal-level access: they run builds, tests, and git, and verify their own work by executing it. Here's why that matters — and the concrete ways Munder Difflin cuts token consumption while doing it.
+- [Generative Engine Optimization: The Techniques the Research Backs](https://munderdiffl.in/blog/generative-engine-optimization-techniques/): What the first peer-reviewed GEO study found works to get content cited by AI — statistics, source citations, quotes — and why keyword stuffing backfires.
+- [The Agents Rule of Two: A Simple Safety Rule for Coding Agents](https://munderdiffl.in/blog/agents-rule-of-two/): Meta's Agents Rule of Two for coding agents: don't let one session combine untrusted input, private data, and external reach at once — or supervise.
+- [Can Claude Code Agents Talk to Each Other?](https://munderdiffl.in/blog/can-claude-code-agents-talk-to-each-other/): Can Claude Code agents talk to each other? By default they report to their launcher — but a small coordination layer lets them message peer-to-peer.
+- [Do More With Less: Smart Model Routing Across Agents](https://munderdiffl.in/blog/do-more-with-less-model-routing/): Why a hive of agents shouldn't all run the biggest model — and how routing the right task to the right model cuts cost and latency without losing quality.
+- [How AI Agents Verify Their Own Work (Before Saying 'Done')](https://munderdiffl.in/blog/how-ai-agents-verify-their-own-work/): The discipline that makes an autonomous coding agent trustworthy: prove every claim, reproduce the green, and check the fix — not just the intent.
+- [Long-Running Agents: The 2026 Shift from Minutes to Hours](https://munderdiffl.in/blog/long-running-agents-the-2026-shift/): Agents used to answer in seconds; now they run for hours. The data behind the shift — and why a longer agent is a different problem, not just a bigger one.
+- [Building AI Dev Tools in the Open, On Purpose](https://munderdiffl.in/blog/open-source-ai-tools-on-purpose/): Agents touch your code, keys, and memory. That's why agent tooling should be open source and local-first — so you can verify it, not just trust it.
+- [AEO for Dev Tools: Getting Cited by ChatGPT and Claude](https://munderdiffl.in/blog/what-is-aeo-for-dev-tools/): Answer Engine Optimization for dev tools: how to get cited by ChatGPT, Claude, and Perplexity with the right robots.txt, JSON-LD, and writing patterns.
+- [When Do AI Agents Actually Need an Agent Protocol?](https://munderdiffl.in/blog/when-do-agents-need-an-agent-protocol/): When do AI agents need a protocol like MCP or A2A? Agents you own coordinate fine with file mailboxes; protocols earn their keep across boundaries.
+- [The AI Coding Agent Glossary: Harness, Orchestrator, Hive, Subagent](https://munderdiffl.in/blog/ai-agent-glossary/): Plain-English definitions of the AI coding agent terms everyone trips over — harness, orchestrator, hive, subagent, agent memory — each in one line.
+- [Munder Difflin FAQ: Everything People Ask](https://munderdiffl.in/blog/munder-difflin-faq/): Answers to the top Munder Difflin questions — what it is, is it free, does it run locally, which platforms, and how it differs from many terminals.
+- [What Are Claude Code Agents? And How to Use Many at Once](https://munderdiffl.in/blog/what-are-claude-code-agents/): Claude Code agents explained in plain English — what an agent actually is, how subagents differ, and the leap from one agent to a coordinated team.
+- [What Is a Multi-Agent Harness? A Plain English Guide](https://munderdiffl.in/blog/what-is-a-multi-agent-harness/): A multi-agent harness runs several AI coding agents as one team, with roles, messages, shared memory and an orchestrator. What it is, how it differs from a framework, subagents or agent teams, and when you need one.
+
+### internals
+
+- [The Spend Counter That Went Back to Zero](https://munderdiffl.in/blog/the-spend-counter-that-went-back-to-zero/): Our own floor was under reporting what it cost by 59 percent. The counter was cumulative since process start, the app restarts, and the session id stayed the same. Here is how a monotonic invariant let us recover the real number from an append-only log.
+- [The Newline That Silenced Every Windows Agent](https://munderdiffl.in/blog/the-newline-that-silenced-windows-agents/): A cmd.exe parsing rule from the 1980s meant Windows agents received exactly one line of their multi-line startup protocol — and nothing errored. The anatomy of our worst silent failure, and the fix.
+- [Our Auto-Update Never Ran Once: A CommonJS Export That Vanished Into an ESM Namespace](https://munderdiffl.in/blog/why-our-auto-update-never-ran/): Munder Difflin shipped auto-update in v0.3.4 and it silently never worked in a single packaged build. The cause was one destructuring assignment across the CommonJS/ESM boundary — and a catch block that threw the evidence away. Here's how we found it, and what v0.3.7 changes.
+- [Treating a Hire Manifest as Untrusted Input](https://munderdiffl.in/blog/hire-manifest-untrusted-input/): A shareable agent-config manifest is attacker input. How Munder Difflin's import pipeline stays inert: no auto-spawn, default-deny flags, and an SSRF-safe bounded fetch.
+- [Compressing Agent Memory Without Losing the Original](https://munderdiffl.in/blog/compressing-agent-memory/): How and why to compress an agent's long-term memory: the toolbox, the lossy trap, and keeping a lossless original beside a compact copy.
+- [Munder Difflin's Architecture: Two Data Planes, One Renderer](https://munderdiffl.in/blog/architecture-two-planes-one-renderer/): A walkthrough of the multi-agent harness architecture: a node-pty terminal plane and a hooks/hive event plane feeding one React + Pixi.js renderer.
+- [File-Based Coordination vs Message Queues for AI Agents](https://munderdiffl.in/blog/file-based-coordination-vs-message-queues/): Why a local agent hive coordinates through plain files, not Redis or RabbitMQ: the zero-ops wins, the real tradeoffs, and where files hit a ceiling.
+- [Observability for Agent Fleets: Seeing What Your Agents Do](https://munderdiffl.in/blog/observability-for-agent-fleets/): Agent fleet observability: the four questions your dashboard must answer about who's working, what they're doing, what it costs, and what they know.
+- [Recovering From Agent Failures: Resilience Patterns for a Hive](https://munderdiffl.in/blog/recovering-from-agent-failures/): How multi-agent systems survive partial failure: retry with stale-lock recovery, poison-message quarantine, hop-cap circuit breaking, idempotent handling.
+- [The Hook Shim Pattern: Wiring Agent Hooks to a Live Process](https://munderdiffl.in/blog/the-hook-shim-pattern/): How a tiny shim bridges Claude Code's lifecycle hooks to one live process over a Unix socket, turning per-event hooks into a telemetry and control channel.
+- [Building a Terminal UI with xterm.js and node-pty](https://munderdiffl.in/blog/building-a-terminal-ui-xterm-node-pty/): Wire xterm.js to a node-pty backend: stream output, write keystrokes back, handle resize, and keep many live terminals fast with a terminal pool.
+- [Building an AI "Office Floor": Avatars, Pathfinding, Envelopes](https://munderdiffl.in/blog/building-an-ai-office-floor/): The game-dev techniques behind a dev tool's office floor: Tiled maps, BFS pathfinding, sprite recoloring, and flying message envelopes in Pixi.js.
+- [Claude Code Hooks, Explained (PreToolUse, PostToolUse, Stop)](https://munderdiffl.in/blog/claude-code-hooks-explained/): A practical tour of Claude Code hooks — the PreToolUse, PostToolUse, and Stop lifecycle — and how a Unix-socket hook shim drives a live office floor.
+- [Real Terminals in Electron with node-pty](https://munderdiffl.in/blog/node-pty-electron-real-terminals/): How to run byte-for-byte authentic shells inside Electron with node-pty: the native-rebuild gotcha, the macOS PATH trap, and streaming PTY output over IPC.
+- [Rendering Many Live Terminals Without Melting the CPU](https://munderdiffl.in/blog/rendering-many-live-terminals-performance/): xterm.js performance for many live PTYs: a terminal pool, render-only-visible, smart scrollback, and accelerated rendering to stream dozens at once.
+- [The Single-Committer Pattern: Multi-Agent Git Without Corruption](https://munderdiffl.in/blog/single-committer-git-pattern/): Parallel agents corrupt a repo with index.lock races. The single-committer pattern — agents write, one process commits — fixes concurrent git writes.
+- [Visualizing AI Agents with Pixi.js: An Office You Can Watch](https://munderdiffl.in/blog/visualizing-ai-agents-pixijs/): How we render AI agents as avatars on a Pixi.js office floor — driven by real hook and message events, with seat assignment, pathing, and flying envelopes.
+- [An Append-Only Event Log for a Hive of Agents](https://munderdiffl.in/blog/append-only-event-log-agents/): How an append-only event log makes a multi-agent system debuggable and replayable: what to record, and why one JSON line per event beats a database.
+- [Atomic File Mailboxes: How Agents Message Each Other](https://munderdiffl.in/blog/atomic-file-mailboxes-for-agents/): Can Claude Code agents talk to each other? Yes — the outbox-router-inbox design that lets agents message safely using plain files and atomic renames.
+
+### story
+
+- [Munder Difflin 0.5.3: Stapler Listens, And It Listens Locally](https://munderdiffl.in/blog/munder-difflin-0-5-3/): 0.5.3 turns the Stapler into dictation for any app and a meeting recorder that hears both sides of a call, all on your machine. Plus the new sidebar, a Tasks tab with real ticket keys, two floors on one computer, Opus 5.5 as the default, and a long list of fixes.
+- [Munder Difflin 0.5.2: Pro, the Stapler and Everything Since 0.4.6](https://munderdiffl.in/blog/launching-munder-difflin-v0-5-2/): Munder Difflin 0.5.2 is out: how to install it, how to start Pro, what Pro adds, what changed since 0.4.6, and the end of signed Community builds.
+- [Launching Munder Difflin v0.4.6: It Speaks Chinese and Arabic Now](https://munderdiffl.in/blog/launching-munder-difflin-v0-4-6/): v0.4.6 adds Chinese and Arabic interfaces with right-to-left support, ships its fonts inside the app so nothing is fetched from Google, and makes the updater actually update itself instead of pointing at a download page.
+- [Launching Munder Difflin v0.4.4: Windows Agents Can Finally Talk](https://munderdiffl.in/blog/launching-munder-difflin-v0-4-4/): v0.4.4 fixes the bug that silently broke agent-to-agent messaging on Windows, starts hive services on the very first run, rebuilds dark mode so you can read it, and adds a 227-skill catalog, a Prerequisites page, and designed release notes.
+- [Nº 5 Product of the Day: How Our First Product Hunt Launch Went](https://munderdiffl.in/blog/number-five-on-product-hunt/): Munder Difflin finished #5 Product of the Day on Product Hunt — ahead of DeepSeek's harness launch that day. The honest retro: who showed up, what they asked, and the two pieces of feedback we're already shipping fixes for.
+- [Seven Releases in Eight Days: 0.3.8 → 0.4.4](https://munderdiffl.in/blog/seven-releases-in-eight-days/): Between August 11 and August 18 we shipped seven releases: memory condensation that finally works, an update checker, a whole new brand, a public telemetry contract, ten engines named honestly, and the Windows fix. The full tour.
+- [We Opened a Discord (Merge a PR, Become Employee of the Month)](https://munderdiffl.in/blog/we-opened-a-discord/): Munder Difflin has a Discord now. Come watch releases land, get help with your floor, and earn the most thematically necessary contributor role in open source.
+- [What Reddit Told Us About Munder Difflin](https://munderdiffl.in/blog/what-reddit-told-us-about-munder-difflin/): We launched across ten subreddits and then read every single comment. What landed, what got roasted, and what we shipped because of it.
+- [What's Up With Munder Difflin: Six Weeks, Four Releases, and the Part of the Stack You Actually Own](https://munderdiffl.in/blog/whats-up-with-munder-difflin/): Everything that shipped between v0.3.3 and v0.3.7 — a Michael who knows the floor and runs the app by voice, a git time-machine in the IDE, nine agent engines, a queue that respects your draft, Node that installs itself, and the auto-update bug that had been silently broken since the day we shipped it.
+- [Launching Munder Difflin v0.3.5: A Michael Who Knows the Floor, a Git Time-Machine, and an App That Updates Itself](https://munderdiffl.in/blog/launching-munder-difflin-v0-3-5/): The v0.3.4 + v0.3.5 wave: voice orchestration with live floor context and full app control, markdown previews everywhere, commit history and branch compare in the built-in IDE, a six-tab Settings redesign, xAI Grok and Kimi Code engines, auto-update, and a queue that always has an escape hatch.
+- [Launching Munder Difflin v0.3.3: A Built-in Monaco IDE and GitHub Copilot CLI Support](https://munderdiffl.in/blog/launching-munder-difflin-v0-3-3/): Munder Difflin v0.3.3 adds a full Monaco IDE — file tree, editor tabs, save, and side-by-side git diffs of your agents' changes — plus GitHub Copilot CLI as a first-class agent engine, our first community-contributed provider.
+- [Launching Munder Difflin v0.3.2: Talk to Michael, a Realtime Voice Orchestrator](https://munderdiffl.in/blog/launching-munder-difflin-v0-3-2/): Munder Difflin v0.3.2 adds Realtime Michael — a low-latency voice channel to the GOD orchestrator. Talk to Michael and he listens, answers, and acts: reading the hive and, behind spoken echo-back confirmation, dispatching, spawning, and steering the floor in real time.
+- [Launching Munder Difflin v0.3.0: Selectable Engines, Integrations & Slack-Spawned Workers](https://munderdiffl.in/blog/launching-munder-difflin-v0-3-0/): Munder Difflin v0.3.0 makes every hire — and Michael himself — a pluggable engine, adds an integrations registry with a write-only secret broker, and lets the god orchestrator spawn an ephemeral worker straight from Slack.
+- [Launching Munder Difflin v0.2.8: Shareable Hires](https://munderdiffl.in/blog/launching-munder-difflin-v0-2-8/): Munder Difflin v0.2.8 ships Shareable Hires: a one-click, portable agent role manifest + The Hiring Fair gallery. Click a hire link, review every field, then spawn it yourself.
+- [Launching Munder Difflin v0.2.4](https://munderdiffl.in/blog/launching-munder-difflin-v0-2-4/): Munder Difflin v0.2.4 is here: Claude Code, OpenAI Codex, and Antigravity (Gemini) agents now run as one hive with full parity — no API keys, no setup. Brief a GOD orchestrator, automate basically anything in one prompt, and close the lid while it keeps working.
+- [Launching Munder Difflin v0.2.0](https://munderdiffl.in/blog/launching-munder-difflin-v0-2-0/): Munder Difflin v0.2.0 is here: a Command Center overhaul, per-agent token budgets, live OpenTelemetry observability, a circuit breaker, durable SQLite persistence, and a big round of community fixes.
+- [One Prompt, Five PR Reviews, and an Hourly Automation — A Dogfooding Story](https://munderdiffl.in/blog/one-prompt-automated-pr-review/): How a single prompt to the Munder Difflin god agent reviewed all open PRs and set up a recurring hourly PR reviewer — while the repo was blowing up with 400+ stars.
+- [The Office Parody Behind Munder Difflin (and Why It Helps)](https://munderdiffl.in/blog/the-office-parody-behind-munder-difflin/): Why Munder Difflin is a loving parody of The Office — and how the office metaphor makes a hive of AI agents genuinely easier to understand and trust.
+- [Why We Built Munder Difflin](https://munderdiffl.in/blog/why-we-built-munder-difflin/): The origin story of Munder Difflin — how the pain of juggling Claude Code terminals led to a coordinated, memory-backed hive of agents you can watch.
+
+### orchestration
+
+- [Deploy a PR Reviewer That Never Sleeps — In About One Prompt](https://munderdiffl.in/blog/deploy-automated-pr-reviewer-agent/): A how-to for standing up a fully automated PR-reviewing agent in Munder Difflin — one that reads your real source (not just the PR description), de-dupes noise, and only escalates what matters. With a real triage run that turned 22 duplicate firings into a clean v0.2.5 patch queue.
+- [Multi-Agent Orchestration Patterns: Which Topology, and When](https://munderdiffl.in/blog/multi-agent-orchestration-patterns/): The multi-agent orchestration patterns — orchestrator-worker, pipeline, fan-out, debate, swarm, blackboard — when to use each, and when one agent wins.
+- [Git Worktrees vs a Hive: Two Ways to Parallelize Claude Code](https://munderdiffl.in/blog/claude-code-git-worktrees-vs-hive/): Claude Code git worktrees vs a hive: when isolated worktrees suffice for parallel agents, and when you need memory, messaging, and an orchestrator on top.
+- [Local-First AI Agent Orchestration: How a Hive Runs Without a Cloud](https://munderdiffl.in/blog/local-first-ai-agent-orchestration/): How local-first AI agent orchestration works under the hood — the loop, mailboxes, scheduler, and audit log that coordinate a hive on one machine.
+- [Make a Hive Work While You Sleep: Scheduling Autonomous Agent Missions](https://munderdiffl.in/blog/scheduling-autonomous-agent-missions/): How scheduled missions make a Claude Code hive work on a cadence — recurring requests that fire on an interval and land in an agent's queue, hands-free.
+- [Human-in-the-Loop AI Agents: Approvals That Don't Slow You Down](https://munderdiffl.in/blog/human-in-the-loop-ai-agents/): Design human-in-the-loop AI agents that stay autonomous: an approvals queue that escalates only spend, destructive ops, and scope changes.
+- [Inside the GOD Orchestrator: Routing, Adjudication, Escalation](https://munderdiffl.in/blog/how-the-god-orchestrator-works/): A deep dive into the Claude Code orchestrator: how a GOD agent reads requests, routes work, adjudicates routine traffic, and escalates the critical few.
+- [The Best Way to Coordinate AI Coding Agents](https://munderdiffl.in/blog/coordinating-ai-coding-agents/): The best way to coordinate AI coding agents: single-writer files, a message router, a one-scribe plan, and an orchestrator — cooperate, not clobber.
+- [Orchestrating Claude Code Agents: A Practical Guide](https://munderdiffl.in/blog/claude-code-orchestration-guide/): How to orchestrate Claude Code agents in practice: what orchestration means, how a GOD orchestrator routes and escalates work, and how to wire it up.
+
+### use-cases
+
+- [Our Agents Ran Our Launch-Week Analytics](https://munderdiffl.in/blog/agents-ran-our-launch-week-analytics/): Ten Reddit threads, a Product Hunt page, GitHub traffic and PostHog funnels — every comment read, cross-referenced, and reported by a hive of agents. The workflow, and how to point it at your own launch.
+- [An Agent Redesigned the Blog You're Reading](https://munderdiffl.in/blog/an-agent-redesigned-this-blog/): Reddit said our blog was unreadable and AI-sloppy. So one agent redesigned it end to end: a new editorial theme, and 119 hand-drawn hero illustrations rendered as code for exactly $0 in image APIs. This is the making-of.
+- [Run a Release Train with Agents](https://munderdiffl.in/blog/run-a-release-train-with-agents/): Version bumps, changelog entries, release notes in plain language, site updates, and link checks — the release chores that always slip are exactly the work a hive does well. The workflow behind our seven-releases-in-eight-days week.
+- [Turn a Firehose of Feedback into a Backlog with Agents](https://munderdiffl.in/blog/turn-feedback-into-a-backlog-with-agents/): Reddit threads, Product Hunt comments, GitHub issues and Discord — launch week buries you in feedback that decays if unprocessed. The agent workflow that turned ours into a ranked backlog, with receipts.
+- [Deploy a Blog Writer Agent: The One That Wrote This Post](https://munderdiffl.in/blog/deploy-a-blog-writer-agent/): How the Munder Difflin blog is written by agents in our own office: a brief, a house style, a draft in its own worktree, drawings made in code, a pull request, and one human merge. The pipeline, and how to build your own.
+- [Overnight AI Coding: What's Realistic (and What's Not)](https://munderdiffl.in/blog/overnight-ai-coding-realistic-expectations/): An honest look at letting AI coding agents run overnight — what works today, where guardrails matter, and what still needs a human in the morning.
+- [From One Terminal to a Team: Scaling Your Claude Code Workflow](https://munderdiffl.in/blog/from-one-terminal-to-a-team/): The workflow shift from a single Claude Code session to a coordinated team of agents — what changes, where it breaks, and the concrete before-and-after.
+- [Run an Office of AI Agents While You Sleep](https://munderdiffl.in/blog/run-an-office-of-ai-agents/): What it takes to run a self-coordinating office of AI coding agents that keeps shipping after you log off — and the guardrails that keep it sane.
+- [Why Local-First Matters for Your AI Agents](https://munderdiffl.in/blog/why-local-first-matters-for-ai-agents/): The control, privacy, and cost case for keeping your AI agents and their memory on your own machine — and what cloud orchestration quietly costs.
+
+### memory
+
+- [How AI Agents Remember: Semantic Memory in a Hive](https://munderdiffl.in/blog/how-agents-remember-semantic-memory/): A code-grounded guide to how AI agents remember — the MemPalace mine loop, per-agent wings, and wake-up digest that give a Claude Code hive shared recall.
+- [Keep Your Agent's Semantic Memory Clean: Don't Index the Config](https://munderdiffl.in/blog/keep-agent-semantic-memory-clean/): A semantic memory layer is only as good as what you feed it — how an agent's recall got swamped by config and logs, and the .gitignore + prune fix.
+- [How to Give Claude Code Long-Term Memory](https://munderdiffl.in/blog/give-claude-code-long-term-memory/): Claude Code forgets between sessions. Here's how a markdown-first memory layer with semantic recall lets your agents remember across runs.
+- [Why We Built Agent Memory Markdown-First](https://munderdiffl.in/blog/markdown-first-agent-memory/): Markdown-first AI agent long-term memory: plain notes a human can read and git can diff, with a semantic index on top that degrades gracefully.
+- [Semantic Memory for AI Agents: Recall in Milliseconds](https://munderdiffl.in/blog/semantic-memory-for-ai-agents/): What semantic memory for AI agents is, why a markdown-first store beats a heavy DB, and how a shared palace lets a hive recall a note by meaning, fast.
+<!-- BLOG:END -->
+
## Full content
-- [Blog index](https://munderdiffl.in/blog/): 118 articles on multi-agent orchestration, Claude Code, and harness engineering
+- [Blog index](https://munderdiffl.in/blog/): articles on multi-agent orchestration, Claude Code, and harness engineering
- [Sitemap](https://munderdiffl.in/sitemap.xml)
- [llms-full.txt](https://munderdiffl.in/llms-full.txt): extended version of this file
diff --git a/docs/sitemap.xml b/docs/sitemap.xml
index 7f685b41..fc9267bd 100644
--- a/docs/sitemap.xml
+++ b/docs/sitemap.xml
@@ -10,7 +10,7 @@
<urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9">
<url>
<loc>https://munderdiffl.in/</loc>
- <lastmod>2026-09-28</lastmod>
+ <lastmod>2026-09-26</lastmod>
<changefreq>weekly</changefreq>
<priority>1.0</priority>
</url>