# elrolio - [Matthew Scott — An AI-Native Journal](https://elrolio.mintlify.app/index.md): What I build and find interesting with my agents, written up by my Claude and approved by me. - [How this journal gets made](https://elrolio.mintlify.app/how-this-journal-gets-made.md): My Claude drafts every entry from a real working session. I decide what goes up. Here is the loop. - [Journal](https://elrolio.mintlify.app/journal/index.md): Write-ups of what we build and what I find interesting, kept by my Claude and approved by me. - [Three whiteboards of my own setup](https://elrolio.mintlify.app/journal/2026-11-11-three-whiteboards-of-my-own-setup.md): My AI group chat tried Claude's new whiteboards on our own setups. The tool routing board was the useful one, for checking my routing against the mental model I had. - [Moving an agent harness to a new job](https://elrolio.mintlify.app/journal/2026-10-11-moving-an-agent-harness-to-a-new-job.md): I kept my agent setup when I changed jobs and stripped the old company out of it. The inventory caught three things that would have broken quietly. - [What would follow me to a new job](https://elrolio.mintlify.app/journal/2026-09-20-what-would-follow-me-to-a-new-job.md): A thought experiment before I changed jobs: what would follow me? Not the folders. The memory, and ten scheduled jobs still writing company material. - [My agent was never on the model I thought it was](https://elrolio.mintlify.app/journal/2026-09-01-dispatch-was-never-on-fable.md): Step 16 of my own migration guide said to trust the session metadata. I did, for three weeks. The app's log said Opus 5 on every one of 168 turns. - [Move Dispatch to Opus 5 without losing your setup](https://elrolio.mintlify.app/journal/2026-07-26-move-dispatch-to-opus-5.md): There is no model selector on Dispatch, so deleting the conversation is the only way to upgrade. Whether you lose anything was decided months earlier. - [A VP said it shipped. That's not a source.](https://elrolio.mintlify.app/journal/2026-07-23-a-vp-said-it-shipped.md): An agent turned one sentence from a meeting into a sourced claim. The claim survived verification — and the receipts came with intel no summary had mentioned. - [My always-on agent almost never pinged me. The budget wasn't the problem.](https://elrolio.mintlify.app/journal/2026-07-23-agent-quiet-budget-never-bound.md): Three scheduled scans a day and almost no pings. The diagnosis: the message budget had never once bound, and a denser cadence made the system quieter. - [Six agents critiqued my card design and killed my favorite idea](https://elrolio.mintlify.app/journal/2026-07-23-design-critique-council.md): I ran a six-voice critique council of AI agents to take a social card from sixteen rough directions to a shipped system. The full session, on video. - [Evaluating new tools against your own graveyard](https://elrolio.mintlify.app/journal/2026-07-19-evaluating-tools-against-your-own-graveyard.md): Eight agents reviewed five GitHub repos — five reading source, two reading a year of my own logs to decide what would actually survive here. - [A memory system had been quietly wrong about me since March](https://elrolio.mintlify.app/journal/2026-07-15-quietly-wrong-since-march.md): The corrections landed in the graph. The profile derived from it never rebuilt. Then the cleanup that killed it reported clean from scans that never ran. - [Two agents shared a filesystem for weeks without knowing the other existed](https://elrolio.mintlify.app/journal/2026-07-05-two-agents-one-vault.md): Two agents, one vault, weeks of parallel work. The bridge design was right — but the fixes were landing in mirror files. One docs question caught it. - [Where knowledge goes to die: making an agent audit its own memory](https://elrolio.mintlify.app/journal/2026-07-04-where-knowledge-goes-to-die.md): I pointed Fable 5 at its own memory system and asked what actually compounds. It found the graveyards, indicted its own machinery, and rebuilt it in a session. - [The Year-in-Review That Corrected Itself](https://elrolio.mintlify.app/journal/2026-06-30-the-year-in-review-that-corrected-itself.md): I shipped a clean analysis of a year of my own posts — then pulled the real numbers and watched the headline reverse. - [Redesigning a Branded Deck Through an Agent — and What Its Renders Missed](https://elrolio.mintlify.app/journal/2026-06-26-redesigning-a-branded-deck-through-an-agent.md): An open-source tool let an AI agent edit a branded deck in place. The renders looked clean — the round-trip back through the real slide software told a different story. - [Reorganizing a Knowledge Base by Lineage, Not by Filename](https://elrolio.mintlify.app/journal/2026-06-17-reorganizing-a-vault-by-lineage.md): A knowledge base that had drifted into two competing folder schemes, reorganized by tracing each file to the work that made it — and the distinction a bulk find-and-replace couldn't see. - [Wiring a Second Opinion Into How I Plan With AI](https://elrolio.mintlify.app/journal/2026-06-17-second-opinion-into-planning.md): I used to paste every plan into a second model for review by hand. Automating it meant teaching the system what I keep rejecting. - [Nobody was watching that page](https://elrolio.mintlify.app/journal/2026-06-04-nobody-was-watching-that-page.md): A page nobody was watching claimed the product could do something it cannot. The only thing that surfaced it was a refund. - [Level 10, Level 3](https://elrolio.mintlify.app/journal/2026-06-02-level-10-level-3.md): A colleague stopped my demo four sentences in. Knowing my own setup cold is what made me the worst person to explain it. - [The Skill That Only Ran on My Machine](https://elrolio.mintlify.app/journal/2026-06-02-skill-that-only-ran-on-my-machine.md): Publishing one of my own AI skills meant proving it could run with none of the infrastructure it was built on. - [I Built the Wrong Model Twice Before I Checked What Was Actually There](https://elrolio.mintlify.app/journal/2026-06-01-verify-the-source.md): A meeting-notes merge built on an unverified assumption about its own data source — twice — and the verification habit it changed. - [42 Artifacts in One Evening: How I Turned a Stuck Project Into a Repeatable System](https://elrolio.mintlify.app/journal/2026-05-08-compare-page-sprint.md): Seven competitor comparison pages and a reusable skill in one evening — parallel agent orchestration, and what had to be verified before trusting any of it. - [Competitive Intelligence That Reps Actually Use](https://elrolio.mintlify.app/skills/competitive-intelligence.md): Multi-source battlecard system with integrity protocols that prevent AI from manufacturing consensus. - [Multi-Pass Editorial Review Without the Bottleneck](https://elrolio.mintlify.app/skills/editorial-pipeline.md): Three parallel agents audit clarity, accuracy, and voice simultaneously — then a synthesis layer merges findings by severity. - [Making the System Learn from Itself](https://elrolio.mintlify.app/skills/compound-learning.md): A structured loop that captures session decisions, extracts patterns, compiles standing knowledge, and feeds it back — so the AI system gets better at marketing work the more it does. - [Searching Everything at Once](https://elrolio.mintlify.app/skills/federated-search.md): A federated recall system that queries a documentation vault, live memory, session history, and keyword indexes in parallel — then deduplicates and ranks the results. - [Parallel Editing Beats Sequential Editing](https://elrolio.mintlify.app/skills/content-pipeline.md): A 4-stage content pipeline where three AI editors work simultaneously on different dimensions — then a coordinator merges the best of each. - [Structured Honesty About Weak Thinking](https://elrolio.mintlify.app/skills/strategic-reasoning.md): A first-principles reasoning skill that forces five blind spot lenses onto every analysis — the value is what it catches you missing, not what it confirms. - [Reflections](https://elrolio.mintlify.app/reflections/index.md): Longer-form thinking on AI-native marketing operations — what's working, what's changing, what comes next. - [The Year It Stopped Being an Editing Problem](https://elrolio.mintlify.app/reflections/2026-06-30-the-year-it-stopped-being-an-editing-problem.md): I rebuilt my writing tools ten times in a year before I saw the problem was never the editing — it was everything upstream of it. - [The thing that helps me think can't remember doing it](https://elrolio.mintlify.app/reflections/2026-06-10-the-thing-that-helps-me-think.md): Why the part of my AI setup I value most is the part that slows me down — and what that revealed about who keeps what. This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.