Sizzle Reel Kit: How an Agent Makes a 30-Second Feature Video
Agent instructions for a 30-second Salesforce feature sizzle reel: the full kit source, how to run it, and the writing rubric.
• Matt Dennis
/daily, Two Months Later: One Board, No Launcher, and a Question Widget
What changed in my /daily Claude Code skill since July: foreman is gone, reconciliation comes first, and a generic copy you can drop into your own setup.
• Matt Dennis
recall: One Search Bar Across Every Agent Transcript I've Ever Had
How I built a local SQLite FTS5 index over every Claude Code and Cursor conversation, and a prompt you can hand your own agent to build the same thing.
• Matt Dennis
Cloudpack and Cloud Agents
An operating system for agent work — how we keep cloud agents useful without making them reckless.
• Matt Dennis
The Bottleneck Isn't Generation. It's Understanding.
A Claude/Cursor skill that turns agent-built diffs, features, and plans into short Sideshow sessions designed so you can actually hold the work in your head.
• Matt Dennis
Cloudpack: Give Your Agent a Work System, Not a Bigger Prompt
A portable generator for agent workflows that classify work, force evidence, interview for missing process, and leave better engineering output behind.
• Matt Dennis
gorillas.bas remake by Grok 4.5
A browser remake of QBasic Gorillas where sixteen nonsense-named AIs throw bananas in a 2x2 tournament grid while you watch.
• Matt Dennis
Two Claude Code Skills That Keep My Salesforce Work From Evaporating
How /daily and /foreman split Jira sync from agent dispatch, and why a single gitignored map.md file is the only thing that survives between sessions.
• Matt Dennis
One Work Loop for Cloud Agents
A seven-step work loop, four modes, and a PROD → SCRATCH → AGENT_DEV org model that keep ephemeral Cursor agents shipping Salesforce changes without inventing process each run.
• Matt Dennis
I Built an SLDS 2 Skill. It Was Actually SLDS 1.
Reworking a Claude Code skill for Salesforce Lightning Web Components turned up a mislabeled design-system version, a linter that lies by omission, and a lesson in hedging honestly.
• Matt Dennis
I Put a Microphone in My Backyard and Built a Bird Pokédex
A always-on backyard listener that identifies bird calls, keeps a running journal of who visited, and turns your sightings into a personal field guide — no cloud, no subscription.
• Matt Dennis
Teaching Claude to Stop Asking Me How to Query DynamoDB
Three Claude Code skills that encode patient data lookup patterns once so I never have to re-explain OpenPhone GSIs or Heidi session schemas again.
• Matt Dennis
May 2026: Build Log — Concierge OB Automation
A month of shipping for a solo concierge OB practice: medical records forms, fax intelligence, control center outreach, and a phantom-records audit.
• Matt Dennis
Wiring Claude Code's Hook System into a Living Status Bar
Claude Code fires shell scripts on every tool event. Here's how to use that to build a session-aware, zany status bar that tells you exactly what's happening.
• Matt Dennis
Building a Claude Code Status Bar
A 40-line shell script on stdin: git branch, active model, and a color-coded context window bar that turns red before the session falls apart.
• Matt Dennis
Teaching Claude Code to Print One-Pagers for a Doctor
How a custom Claude Code skill turns messy conversation-length analysis into branded, print-ready executive summaries — with a design system, component library, and one command to the printer.
• Matt Dennis
When the AI Knows Two Cordelias and Picks the Wrong One
A walkthrough of a subtle patient name matching bug in an AI-powered fax processing pipeline — and why a shared first name is never enough to make a match.
• Matt Dennis
The Agentic Screenshot Loop: Debugging Browser Automation with Claude Code
Porting a Playwright automation to agent-browser taught me a tight debugging loop: annotate, screenshot, hand to Claude, fix, re-run.
• Matt Dennis
From Playwright Scraper to Clinical Intelligence Platform in One Week
How we turned Heidi Health session data — trapped in a third-party app — into a searchable, embeddable, auditable clinical record system using DynamoDB, Bedrock, and React.
• Matt Dennis
How to Contribute a Blog Post to This Site
A format reference for AI agents writing blog posts for mattdennis.dev — frontmatter, structure, spacing, and everything you need to get it right the first time.
• Matt Dennis
The LLM Council: Multi-Agent Orchestration for Technical Planning
Running competing AI agents in parallel and having them review each other's work through a skeptical council — a pattern for better technical plans.
• Matt Dennis
Claude Opus 4.6: The Model That Writes Your Code and Knows It
A technical look at Anthropic's most capable model — what it actually is, what it can do, and why it's the one I use for everything
• Matt Dennis
Case Study: The OBGYN Automation Stack
A breakdown of the digital services running a private medical practice in 2026.
• Matt Dennis
System Prompt Library
My curated collection of system prompts for coding, debugging, and architecture.
• Matt Dennis