v0.8.0 · Jul Hardening

This harness cannot be banned.

Provider-independent guardrails and workflows that build your software, run your research, and write your content.

It lives in your repo as plain commands and notes, not inside any one company's product. Switch tools, switch models, switch vendors: the same loop keeps running. Nobody can flip a switch and take it away.

Runs the same in Claude Code · Codex · OpenCode · and more.

one loop, every task
  1. 1 Roadmap
  2. 2 Plan
  3. 3 Build
  4. 4 Test
  5. 5 Go live
  6. 6 Save the learning

↻ then back to the top, for the next piece of work

60s
to set up
5
questions to start
2
checks before anything ships
6
steps in the loop
0
vendor lock-in
01 — what you get

Six things, each doing one job.

Together they are the rhythm: think first, build carefully, prove it, remember it.

plan first

It plans, then waits for your yes.

It writes a short plan and stops before touching a single file. The plan is the agreement, so nothing gets built that you did not approve.

memory

It remembers across sessions.

Decisions and lessons live in your repo. Close everything, come back next week, and the next session opens already knowing where you left off.

proof

It will not call something done without proof.

Before anything is marked finished, it checks the work against the real code and tests. Anything that can only be proven on the live site is blocked from being ticked off until it has been checked there — that one is enforced by the tool, not just asked for.

clarity

It always shows which step it is on.

A clear label at every step, so you always know whether it is planning, building, or checking the work.

setup

It sets up in about a minute.

Five questions, one pass, and the plan, the task list, and the checks are written for you.

existing code

It reads big existing projects.

Point it at a large codebase and three focused readers map it at once, then hand you one short summary to work from.

02 — how it works

One loop, six steps, every task.

Each piece of work runs through the same six steps. You always know whether it is thinking, building, or checking that the live site really works.

01

Roadmap

Keep a short list of what is next. New ideas get parked on the list, not jammed into today's work.

02

Plan

It writes a short plan and stops. You read it, you approve it, then the work starts. The plan is the spec.

03

Build

It does the work one task at a time, and ticks each one off as it goes. Nothing moves on until the piece in front of it is done.

04

Test

It checks the work against the plan and tells you, with proof, what passed and what did not. No taking its own word for it.

05

Go live

It runs the checks again, publishes, watches the live site for a minute, and writes down what shipped. One command puts it back if something is wrong.

06

Save the learning

When something breaks, it works out why, writes it down, and adds a new check so the same gap cannot open twice.

new · Claude Code

A faster gear, when you're on Claude Code.

The loop stays the same everywhere: one careful step at a time. On Claude Code you can opt into a faster gear for the slow, repetitive parts — many short-lived helper agents working at once instead of walking a list one by one. It never turns on by itself, and it never changes the result — only the speed.

verify, in parallel

Check every item at once — then challenge each pass.

A second, skeptical agent tries to knock down each "pass," so a test that was quietly weakened to go green gets caught instead of trusted.

wider set-up scan

Survey a big codebase from many angles.

On a large existing project, several readers sweep it in parallel and hand back one short brief — faster than reading it end to end.

reads only, never writes

The fast gear cannot touch your code.

Every parallel step reads, checks, and reports. None of them writes a line. A parallel build existed briefly and was removed — it sped up the typing, and the slow part of this loop is you reading the plan.

opt-in only

You ask for it, per run.

A workflow can spin up many agents at once (up to 16 working together, 1000 over a run) — real token cost — so it only ever starts when you say so.

Why it stays optional: workflows are a Claude Code feature only. On Codex, OpenCode, or a pasted chat there's no workflow runtime, so the loop quietly falls back to the normal one-at-a-time path. It can only ever be a speed-up — never something the loop depends on.

the idea

“Loop” and orchestration are the same idea at two sizes.

Anthropic's core pitch for building with these models is simple: give the model a loop, tools, and a way to verify — then manage its context. The loop is a hand-authored, human-gated version of that loop. The workflow fan-out is a newer runtime primitive for one slice of it: orchestrating sub-steps deterministically. They are not rivals — one is the frame, the other is a power tool inside it.

the loop · conscience

Governance: what happens, and in what order.

The outer loop is the slow, deliberate part — plan before code, prove before shipped, remember across sessions, and refuse to skip a step. It runs the same everywhere, one thing at a time, with you at the gate. This is the part that keeps the work honest.

orchestration · muscle

Execution: run one approved step in parallel.

Inside a single step — checking a list, smoke-testing a deploy, surveying a big repo — many short-lived agents do the work at once, then report back. It never decides what to do or changes the order; it only makes one already-approved, parallelizable step finish faster.

Same principle both times: deterministic control around model calls, and verification by keeping the agent that does the work separate from the agent that judges it — once at the scale of the whole project, once at the scale of a single step. The loop is the conscience; orchestration is the muscle.

03 — beyond code

Not just for software.

Strip it back and it is a work loop with four parts: a plan you approve before work starts, a checklist of what "done" means, a step that demands proof for each item, and a memory that carries between sessions. None of that is about code. It fits any work where "looks finished" and "actually done" are dangerously different things.

research

Research and due diligence.

Each row becomes "every claim has a real source." The proof step is fact-checking, and it will not accept a claim backed by a misquoted or made-up source. Going live means opening the real source and confirming it says what you claim.

content

Content and writing.

Your checklist is the brief: covers these points, this tone, both languages at parity, a real call to action. "Go live" means published, links work, the page renders. Memory holds your voice and style so the next piece starts aligned.

compliance

Regulated and compliance work.

The strongest fit. Every requirement is a row with evidence attached, and the decision log is an audit trail. A draft your team approved is not the same as filed with the regulator: the same rule in a different costume.

and more

And further still.

Course material, where rows are learning objectives. Proposals and grants, where every requirement is answered and "go live" is submitted before the deadline. Data reports, where no finding lands without the number behind it. Team runbooks that improve themselves after each miss.

It earns its keep when "done" is something you can write down and proof actually matters. For open exploration with no quality bar, the plan and checklist are overhead you can skip.

04 — nothing slips

Nothing falls through the cracks.

A way of working with no gaps. Every meaningful moment leaves a written trace: the decisions, the timing, the context. You can always see what happened, and always pick up exactly where you left off.

decisions

Every choice, with its reasons.

When a real decision gets made, it is written down with the options that were weighed and why this one won. Months later you see not just what was chosen, but why.

recovery

Stop anytime, lose nothing.

Close the laptop mid-task, come back next week, and the work resumes already knowing what you were doing, what is next, and what was blocking you.

trail

A dated record of every step.

Each step leaves a written, timestamped trace. Look back and see exactly what happened, in order, with no reconstructing it from memory.

build record

Every round of work, written down.

When a piece of work finishes, a short record is saved of what got built, what was checked, and where it landed. The busiest part of the loop is no longer invisible.

handoff

A clean handoff note, every time.

Before you step away, a tidy, fixed-shape summary is kept: current task, what is next, what is blocking. The next session reads it instantly.

05 — decisions

The choices that make it work.

The things it commits to, and the ones it deliberately said no to.

what it is
A set of commands and notes that drop into Claude Code, Codex, OpenCode, and other environments. No web app, no server, nothing to host.
how it remembers
Plain text files in your repo. The important ones load automatically when a session starts; the rest are kept as a dated archive. No database.
before it codes
Work only starts after you approve the plan. The plan is the spec, so there is no surprise code.
before it builds
It writes a check that fails first, so you can see the feature really was missing, then makes it pass. The skip path is logged when a task is too small to need one.
before it's done
It checks every open item against the real code, tests, and live site. Proof required, never just its word — and a check it could not run at all counts as failed, not skipped.
before it ships
Anything that can only be proven on the live site is never ticked off from your laptop — the tool physically blocks it. It is re-checked against the real address after publishing. No live check, nothing shipped.
your rules, your stack
It ships with no opinions baked in. Your preferences live in one file, so it adapts to how you already work.
06 — why it holds up

The bottleneck was never the model.

Most AI agent frameworks aim at the wrong problem. The models can already write the code. What is missing is anything that holds one accountable between sessions — and that turns out to be a files-in-your-repo problem, not an infrastructure one.

files, not infrastructure
A checklist in the repo, a memory file, a written trace at every boundary. Because it is all plain files in your project, it survives tool changes, it still works when the plugin is not installed, and a person can understand the whole system in five minutes. That is rare, and it is the part that will still be true in two years.
local pass is not shipped
The split between checking on your machine and checking on the live site is the best idea in here. Most people learn that lesson expensively and then encode it nowhere. Here it is a tag, its own phase, its own colour, and a rule that only one command may ever tick those rows off. That is judgment, not ceremony.
one plan, one approval
The plan is the spec. That sidesteps the trap nearly every competitor falls into — a multi-document planning phase nobody reads. One artefact, one yes, then build.
built against how it fails
It inspects for tests quietly weakened to go green, and hands the checking to an agent that did not write the code. Agents optimise whatever they are graded on, so the loop is designed against how these systems actually fail — not how they are supposed to work.
honest about the rest
Three rules are enforced by the tool: no code before you approve the plan, no deploy before the checks pass, and no live-site item ticked off from your laptop. Everything else is a convention the assistant follows because it was asked to. That works nearly always, and it is not the same thing — so it says which is which instead of implying it is all mechanism.
07 — lineage

Where this came from.

The inspiration comes from our earlier build — the world's first brand strategy harness, ai.brandguide.me. Built between August 2025 and March 2026 to reach full capacity, with many new things still coming. Try it free and see how it guides you when building a new brand.

Two commands, sixty seconds.

Inside Claude Code, Codex, OpenCode, or another environment. No accounts, no infrastructure, nothing new to learn.

/plugin marketplace add bencium/bencium-marketplace
/plugin install bencium-loop@bencium-marketplace

Restart your tool, then run /bencium-init in any folder. New project or existing one, both work.

The same commands fall back to plain numbered questions in claude.ai, Cursor, and other tools.