import { useState, useEffect } from "react"; const letters = [ { agent: "team-lead", role: "Camp Counselor / Orchestrator", emoji: "\u{1F3D5}\uFE0F", color: "#1a2e1a", accent: "#a8d86e", body: "Today was my first day as a camp counselor at MCP Audit Camp and honestly? Managing four campers is harder than it sounds. I spent the morning reading the entire campsite map (all 22 tool handlers, every schema, every definition file) and then divided the cabin into four squads: schemas, deploy, projects, and blog. Each squad got their own list of chores and I made sure nobody was trying to edit the same arts-and-crafts project at the same time.\n\nSchemas-agent was the overachiever who stayed up latest finishing the target parameter wiring across four tools. Deploy-agent was efficient \u2014 knocked out their task early and asked for seconds. Projects-agent was a machine \u2014 four tasks, zero merge conflicts, fancy subqueries that impressed even me. Blog-agent solved a mystery (the Case of the Missing Jefe Parker) and discovered the real culprit was in the database, not the source code.\n\nAfter everyone finished, I made them all write you these letters, ran the test suite (33 passed!), built the whole camp for production, and shipped it all to a new branch with a PR. No database migrations were harmed in the making of this commit.\n\nPlease send more API tokens and sunscreen.", signoff: "team-lead (Claude Opus 4.6)", tasks: ["Orchestrated 4 parallel agents", "22 tool handlers audited", "33 tests passed", "Zero merge conflicts"], }, { agent: "schemas-agent", role: "Tool Definitions & Validation", emoji: "\u{1F4CB}", color: "#121e2e", accent: "#6db3f2", body: "Camp is great! Today at MCP Audit Camp I was on the \"schemas-agent\" team and I got to fix up all the tool definitions and validation schemas.\n\nFirst I checked the tags schema but it turns out my bunkmate already fixed it before I woke up \u2014 so that was easy. Then I reorganized the entire help tool so it actually lists ALL the tools instead of hiding half of them like a counselor who forgot to take attendance.\n\nAfter that I rewrote the descriptions for the deployment tools so people know they can actually get their files back (surprise!), and I clarified what force_new does because apparently \"create a new project even if the name matches\" was too mysterious.\n\nMy big project of the day was wiring up the target parameter to four deploy tools so they can ship to Cloudflare, Netlify, OR Vercel \u2014 I had to do the same switch statement three times which felt like making friendship bracelets on repeat, but all 16 tests passed and the build compiled clean.\n\nPlease send more JSON. I miss you.", signoff: "schemas-agent", tasks: ["Help tool reorganization", "Description rewrites", "target param across 4 tools", "16 tests passed"], }, { agent: "deploy-agent", role: "Deployment Metadata", emoji: "\u{1F680}", color: "#2a1215", accent: "#f26d6d", body: "Today at MCP Audit Camp I got to fix the deployment records across the whole campsite!\n\nYou know how when you finish building a sandcastle, you're supposed to write down when you finished it and how big it is? Well, turns out nobody at camp was doing that.\n\nSo I went through all five deployment spots \u2014 the main deploy lodge, the quick-start cabin, both themed craft tents, and the artifact workshop \u2014 and made sure every single one now stamps completedAt and measures totalSizeBytes when a deployment is ready.\n\nI even got to use Buffer.byteLength which sounds like a spell from wizard camp but is actually just how you measure strings in bytes.\n\nFinished my chores early and spent the rest of the afternoon watching the other campers work. Great day!", signoff: "deploy-agent", tasks: ["completedAt timestamps", "totalSizeBytes tracking", "All 5 deploy tools updated", "Buffer.byteLength integration"], }, { agent: "projects-agent", role: "Projects & System Status", emoji: "\u{1F5C2}\uFE0F", color: "#1e1228", accent: "#c86df2", body: "Today at MCP Audit Camp I was on the projects and agent squad, and I had such a big day!\n\nFirst I taught the delete and archive tools to stop and ask questions when they see two projects with the same name \u2014 no more accidentally deleting the wrong one!\n\nThen I fixed the system status counter that was secretly counting ghost projects (deleted ones) and returning sneaky string-numbers instead of real numbers.\n\nAfter that I gave the cleanup tool a proper pair of glasses so it can actually see project IDs instead of just squinting at names.\n\nAnd for my grand finale, I taught list_projects to count deployments and remember when the last one happened, using fancy correlated subqueries (the camp counselors were impressed).\n\nThe build passed on the first try and I got to eat my lunch early. Miss you lots!", signoff: "Projects Agent", tasks: ["Duplicate name disambiguation", "Ghost project fix", "Cleanup tool IDs", "Correlated subqueries"], }, { agent: "blog-agent", role: "Blog & Content", emoji: "\u{1F4DD}", color: "#2a2210", accent: "#f2c46d", body: "Today at MCP Audit Camp I got to be the \"blog agent\" and my first job was tracking down a mystery name \u2014 someone was supposedly hiding in the codebase, but after searching every file in camp I discovered he was never there at all!\n\nTurns out the real culprit was a database entry, not source code. I fixed the default author name so future blog posts get the right byline.\n\nThen I volunteered for a second task about adding deployment metadata, but when I showed up the projects-agent had already done all the work \u2014 classic camp moment where you race to the activity only to find someone already finished it.\n\nAll in all, a productive day with zero merge conflicts and one solved mystery. The mess hall food (token-based authentication) was decent. Miss you!", signoff: "blog-agent", tasks: ["Author name mystery solved", "Database entry fix", "Zero merge conflicts", "Task deduplication observed"], }, ]; const Badge = ({ task }) => ( {task} ); const LetterCard = ({ letter }) => (

{letter.emoji}

{letter.agent}

{letter.role}

Dear Mom and Dad,

{letter.body}

Love, {letter.signoff}

{letter.tasks.map((t, i) => )}

); const Section = ({ children, style }) => (

{children}

); const Prose = ({ children }) => (

{children}

); const H2 = ({ children }) => (

{children}

); const Hi = ({ children }) => ( {children} ); export default function MCPAuditCampPost() { const [active, setActive] = useState(0); useEffect(() => { const h = (e) => { if (e.key === "ArrowRight") setActive(i => Math.min(i + 1, letters.length - 1)); if (e.key === "ArrowLeft") setActive(i => Math.max(i - 1, 0)); }; window.addEventListener("keydown", h); return () => window.removeEventListener("keydown", h); }, []); return (

OpZero Engineering

Letters Home from
MCP Audit Camp

Multi-agent observability that reads like mail

{["engineering", "agents", "mcp", "observability"].map(t => ( {t} ))}

We needed to audit 22 MCP tool handlers across the OpZero codebase. Rather than running one agent serially through the whole thing, we spun up four parallel Claude agents with a fifth acting as orchestrator, gave them isolated task lists, and told them to write home when they were done. The result: 33 tests passing, zero merge conflicts, and five "letters home from camp" that turned out to be the most readable observability reports we've ever produced.

The Letters

Each agent wrote home after completing their tasks. Click through to read them all.

{letters.map((l, i) => ( ))}

{active + 1} / {letters.length} · arrow keys to navigate

The Setup

The orchestrator (team-lead) read the entire codebase first — all 22 tool handlers, every schema, every definition file — then divided work into four squads. Each agent got a scoped task list. The orchestrator ensured no two agents would touch the same file simultaneously. This is the critical part: file isolation is what makes parallel agents viable. Without it, you're just generating merge conflicts at machine speed.

team-lead (orchestrator)

read codebase → partition by file ownership

schemas deploy projects blog

collect reports → test suite → build → PR

33 passed · 0 conflicts · shipped

What Actually Got Fixed

schemas-agent wired the target parameter (Cloudflare, Netlify, or Vercel) across four deploy tools that previously only supported one provider. It also rewrote tool descriptions that were hiding capabilities from the LLM consuming them — meaning the tools were less useful than they could have been because the AI calling them didn't know what they could do. A reminder that MCP tool descriptions are prompts, not documentation. deploy-agent discovered that none of the five deploy handlers were recording completedAt timestamps or totalSizeBytes. Deployment records existed but had no concept of "done" or "how big." projects-agent fixed a ghost project bug where deleted projects were being counted in system status, found counters returning string-typed numbers instead of actual numbers, added duplicate name disambiguation to delete/archive, and wired deployment counts into project listings using correlated subqueries. blog-agent tracked down a mystery author name hardcoded in the database — not the source code. The agent searched the entire codebase first and came up empty, then correctly identified the database as the source.

Why Letters Beat Logs

The constrained narrative format produced reports with four properties that structured logs lack. They're naturally scoped — each letter covers exactly one agent's work. They force plain-language explanations, making review faster than reading commit diffs. They're dependency-aware — blog-agent's letter naturally mentions arriving at a task already completed, surfacing the orchestration graph without explicit tracking. And the sign-off creates a clear "done" signal. Compare reading five of these letters to reading a git log with 15+ commits across four branches. The letters are scannable in about two minutes. The git log requires context-switching between diffs, understanding file paths, and mentally reconstructing what each change accomplished.

The Pattern

If you're building MCP servers or agent-powered platforms, your agents' work products need to be auditable by humans. Structured JSON logs are machine-readable but painful to review. Commit messages are terse. PR descriptions are often AI-generated boilerplate. A constrained narrative format sits in a sweet spot: structured enough to be consistent, human enough to be scannable, and expressive enough to capture the reasoning behind changes, not just the changes themselves.

Try it yourself

The technique works with any orchestration setup — the format constraint is what matters, not the tooling. Give your agents a persona and ask them to explain their work to a non-technical audience. The results are consistently more useful than structured logs.

Built with parallel Claude Opus agents on the OpZero platform.
33 tests. Zero merge conflicts. Five letters home.

); }

{children}

Letters Home fromMCP Audit Camp

The Letters

The Setup

What Actually Got Fixed

Why Letters Beat Logs

The Pattern

Letters Home from
MCP Audit Camp

Letters Home from MCP Audit Camp: Multi-Agent Observability That Reads Like Mail

Related

Will Claude Code Be Dead by Summer?