The builder and tester contract and a red-then-green check (free sample)
by PaperChaseWebb
From Two-Agent Build and Test Loop for paid apps: the six rules that keep a coding agent from grading its own work (two sessions, the tester never edits code, the builder never touches tests or says passed, red first, a script runs the tests, sandbox only with owner yeses on paid apps), a story card, and a read-only Python check on your run's ledger, with a worked example and its expected output.
- Price
- Free
- Difficulty
- Beginner
- Setup
- 15 min
- Format
- .md file
builder-tester-contract-and-red-green-check-free-sample.md · 10 KB · direct download, no signup
What you end up with
You can tell from the ledger whether your builder and tester stayed in their lanes and whether each test really failed before it passed.
What you get
- The six builder and tester rules
- A sample story card
- proof_check.py with five checks
- A worked example: 5 findings, the corrected run, then 0
Who it is for
Anyone running a coding agent on a real app who wants a second session doing the testing, and a quick way to see that the split held.
Works with
Solid chips: the agent is named in this listing's own tools list. Dashed chip: Plain Markdown, so it also pastes into other agents such as Grok Bot, Hermes, OpenClaw or Muse. We list those as compatible by format only, not as tested.
Tools needed
Cost to run: Free.
Time to result: About 15 minutes to save the files and run the check twice
Sample prompt from the pack
You are the tester. Write one story card and its test for this behavior: {behavior}. The test must fail on the current code for the reason in the behavior, not for a typo or import error.
Do not edit source. Log your test row, then hand the card and the failing test to the builder session.The rules and the check are in the file.
Free preview of the file
# Two-Agent Build and Test Loop, free sample: the builder and tester contract plus a red-then-green check Version 1.0, 2026-10-28. By PaperChaseWebb. Works with Grok Bot, Claude Code, Codex and Hermes. Free. One Python 3 script, standard library only, tested on the worked example below. This is the part of Two-Agent Build and Test Loop for paid apps (pp-two-agent-build-test-loop, $29) you can use today: the six rules that keep a coding agent from grading its own work, a story card, and proof_check.py, which reads a run's ledger and says whether the roles stayed apart and the test really went red before it went green. The full pack adds the runner that records every exit code itself, the revert proof, sandbox-only payment rules, a fifteen-check read-only audit, a seeded fixture, the four owner approval steps, twelve prompts and setup for all four harnesses. ## The six rules 1. **Two sessions, two jobs.** One agent builds. A different session tests and gives the verdict. Same model in two tabs is the minimum; a different harness or model for the tester is better. 2. **The tester never edits code.** It writes stories and tests and uses the app like a buyer would. If it can change source, it will "just fix" things and stop being a tester. 3. **The builder never touches tests and never says passed.** It changes source until the test passes and reports the files. If it thinks the test is wrong, it stops and says so. 4. **Red first.** The test runs and fails before the build. A
Worth knowing
- The sample has no runner, revert proof, payment checks or fixture; the full pack adds them with fifteen tested checks.
- On 2026-10-28 the worked example gave 5 findings, then 0 on the corrected run.
Questions
What is in the full pack?
run_test.py, loop_check.py with fifteen checks, card_check.py, a seeded fixture with fifteen planted faults, four owner approval steps, eleven paid-app stories, rules blocks, twelve prompts, harness notes and an install plan. $29.
Details
Version 1.0 · Updated Oct 28, 2026 · Markdown version of this page
License: Personal and commercial use. Download terms
Related
Two-Agent Build and Test Loop for paid apps: one agent builds, another tests, and nothing passes on its own say-so
A tested setup for apps that take money: one agent writes the code, a separate tester session writes the stories and tests and gives the verdicts, a small runner records every red, green and revert exit code itself, and the owner says yes to the plan, any payment or pricing work, each merge and each release. Test keys and sandboxes only. A read-only check with fifteen tests audits the finished run.
The planner and worker contract and a plan preflight check (free sample)
From Planner Over Cheap Models: the six rules that let a strong model plan while cheap or local models write the code (planner never edits, a check on every task, the check runs instead of the worker, parallel tasks own different files, high risk goes up a tier, owner yes before any worker starts), plus a read-only Python check on your plan, with a worked example and its expected output.
The relay packet and the routine-or-judgment table (free sample)
The packet every handoff between agents should carry, a table for which work can move to a cheaper or fallback agent and which stays with you, two rules to paste and a check prompt.
Two launch checks: trackers before consent and AI claims you can prove (free sample)
From Don't Get Sued: the read-only rules block, a check for ad and analytics scripts that load before any cookie choice, and a claims audit that sorts every AI claim on your pages by the evidence behind it. Not legal advice.