Monospace · CraftNº 004 · Playbook Edition

The Workflow
That Compounds

Every Inc.'s compound-engineering plugin bets that prompts deserve the same rigor as code — compiled, tested, versioned, and shipped to ten agent platforms from one source.

COMPOUNDING MEMORY
EveryInc/compound-engineering-plugin · MIT · TypeScriptBenched — 1,678 tests + live conversion, June 2026
The Premise02

Really two products: a methodology, and the transpiler that ports it.

The first is 39 skills and 43 agent personas that encode Every's internal loop — strategize, ideate, plan, execute, review, and the step it's named for, compound: write down what was learned so neither humans nor future agents learn it twice. The second is an ~8,900-line TypeScript CLI that ports the whole plugin to a dozen platforms from one source.

21kstars in nine months
39 / 43skills / agent personas
~8,900lines of TypeScript
31Learnings already banked
House split: 80% planning and review, 20% executionThe Workflow That Compounds
The Machine03

A textbook transpiler, with a memory loop underneath.

SourceClaude-native plugin · 39 skills · 43 agent personas
Parseinto one typed intermediate form
Mapper-target mappers · explicit lookup tables for tools, permissions, hooks, model aliases
Writeper-target writers · non-destructive: removed → timestamped backups, configs deep-merged
Memory loopsolved problem → YAML "Learning" → grep-first in later reviews
Mappings are lookup tables, not naming conventionsThe Workflow That Compounds
The Bench04 · Why the plays are believable

One suite, one live conversion, fourteen seconds.

We ran the plugin's own test suite and a live multi-target conversion of the actual plugin. 1,669 of 1,678 tests passed in fourteen seconds — the nine that didn't all need live GitHub access the sandbox blocks, and not one was a real defect. Then a conversion to OpenCode, Codex, and Gemini went through with genuine semantic remapping, not passthrough.

1,669/1,678tests pass, our bench
14sfull suite runtime
3platforms converted live
0.1inferred temp on model: inherit
2runtime dependencies
The nine failures need live GitHub — none was a real defectThe Workflow That Compounds
Play 01Verified An Afternoon

Ship one workflow to ten platforms.

The scenario → the move

Your team is split across Claude Code, Codex, Cursor, OpenCode, and Gemini, and every process doc rots separately for each. Compile the plugin — or your own — to all of them at once and let the lookup tables carry the translation. Installs don't clobber anything.

git clone EveryInc/compound-engineering-plugin
# compile → opencode · codex · gemini · non-destructive
The receipt

A real multi-target conversion of the live plugin succeeded against OpenCode, Codex, and Gemini on our bench — with genuine semantic remapping, not passthrough: a model: inherit persona converted with an inferred temperature of 0.1.

Boundary — the mappers translate, but hand-pinned aliases and targets still drift with the platforms.

The mapper reasons about the target — it doesn't copy bytesThe Workflow That Compounds
Play 02Verified A Weekend

Steal the review panel.

The scenario → the move

You want serious agent-driven code review without adopting anyone else's whole methodology. Take ce-code-review apart and keep three bolts: reviewer selection by judgment, a confidence gate that hides anything below an explicit score, and autofix classes that draw the automation line on purpose.

ce-code-review
# 14 personas · gate at 75 · autofix vs. human-only
The receipt

The orchestration is real in source and the suite backs it: 1,669 of 1,678 tests passed in 14 seconds on our bench — fourteen reviewer personas picked by reading the diff, a default gate at 75.

Boundary — the nine skipped tests need live GitHub the sandbox couldn't give them.

Rebuild three ideas in your harness; skip the whole methodThe Workflow That Compounds
Play 03Demonstrated A Sprint

Bank every solved problem.

The scenario → the move

Your team re-solves the same class of bug every quarter because the fix lives in a closed PR thread. Switch on the Learnings loop, feed it your first ten solved problems, and come back in a quarter with one question: is review starting to cite them?

learnings/*.md   # YAML frontmatter
# retrieved grep-first inside future reviews
The receipt

The loop ships working — the repo carries 31 of its own Learnings and retrieves them during review. Whether the compounding pays off for your team is unproven: the efficacy evidence is testimony from its authors, not measurement.

Boundary — run your own before/after; no one has measured the payoff yet, including Every.

Your before-and-after would be the first hard measurementThe Workflow That Compounds
The Fine Print08

About as clean as scans get. One gap that matters.

Full report: security scan in the study archiveThe Workflow That Compounds
The Ledger09 · The forwardable summary

Three plays, priced by evidence.

PlayTierEffortReceipt
01 Ship one workflow to ten platformsVerifiedAN AFTERNOONlive conversion to OpenCode/Codex/Gemini · real remapping
02 Steal the review panelVerifiedA WEEKEND1,669/1,678 tests in 14s · 14 personas · gate at 75
03 Bank every solved problemDemonstratedA SPRINTloop ships · 31 Learnings retrieved · payoff is testimony
Tiers: Verified — we ran it · Demonstrated — its own loop ran, payoff unmeasuredThe Workflow That Compounds
The Kicker10

Best for an engineering loop that compounds.

If you run one play tonight, run the converter — compile the plugin to every platform your team uses and let the lookup tables carry the translation. It's the first tool we've tested that hands part of the process-fragmentation tax back, down to an inferred temperature on a model: inherit persona.

git clone https://github.com/EveryInc/compound-engineering-plugin
# compile to opencode · codex · gemini — backups + deep-merge
MONOSPACE · monospacemag.comField-tested reviews of trending repositories
← → / space · F fullscreen