Now liveThe Skillselion MCP - thousands of ranked skills, loaded into your agent mid-task. No install.Get it →
broomva avatar

Harness Engineering Playbook

  • 31 installs
  • 25 repo stars
  • Updated June 28, 2026
  • broomva/harness-engineering

Harness Engineering Playbook is a Claude Code skill that implements OpenAI's Harness Engineering practices in a repository, bootstrapping deterministic agent-run commands, docs, observability and CI.

About

This skill operationalizes OpenAI's Harness Engineering practices in a repository so agents can run against it repeatedly and safely. It bootstraps harness artifacts like AGENTS.md, PLANS.md, deterministic smoke/test/lint/typecheck scripts and CI workflows from templates. A developer uses it to make autonomous agent runs reproducible and to enforce architecture boundaries, observability and entropy control.

  • Implements OpenAI's nine Harness Engineering practices in any repo
  • Bootstraps AGENTS.md, PLANS.md, deterministic smoke/test/lint/typecheck scripts and CI
  • Includes an audit command that flags MISSING/FAIL gaps before setup is complete

Harness Engineering Playbook by the numbers

  • 31 all-time installs (skills.sh)
  • Ranked #856 of 1,435 DevOps & CI/CD skills by installs in the Skillselion catalog
  • Data as of Jul 28, 2026 (Skillselion catalog sync)
At a glance

harness-engineering-playbook capabilities & compatibility

Capabilities
repo scaffolding · ci setup · agent tooling · observability setup
Works with
github
Use cases
ci cd · devops · documentation
Pricing
Free
From the docs

What harness-engineering-playbook says it does

Implement OpenAI Harness Engineering practices in any repository.
SKILL.md
Treat any `MISSING` or `FAIL` result as blocking before calling harness setup complete.
SKILL.md
By default, existing files are not overwritten. Pass `--force` to replace template-managed files.
SKILL.md
npx skills add https://github.com/broomva/harness-engineering --skill harness-engineering-playbook

Add your badge

Show developers this skill is listed on Skillselion. Paste this into your README.

Listed on Skillselion
Installs31
repo stars25
Last updatedJune 28, 2026
Repositorybroomva/harness-engineering

What it does

Use it to set up an agent-first repo harness with deterministic commands, docs, observability and CI for reliable autonomous runs.

Who is it for?

Setting up or refactoring agent-first workflows, writing AGENTS.md and PLANS.md, and wiring deterministic harness commands plus CI

Skip if: Application feature work; it operates on repo tooling, docs and CI rather than product logic

When should I use this skill?

Setting up agent-first workflows, creating harness commands, or adding entropy-control checks and CI

What you get

A repo with deterministic harness commands, compact agent docs, observability and CI that pass an audit gate

  • AGENTS.md
  • PLANS.md
  • harness scripts

By the numbers

  • 9 Harness Engineering practices
  • 3 profiles (baseline, control, full)
  • 5-step workflow

Files

.skills/harness-engineering-playbook/SKILL.mdMarkdownGitHub ↗

Harness Engineering Playbook

Use this skill to operationalize the practices from OpenAI's Harness Engineering guide in a repo that agents can run against repeatedly and safely.

What To Load

  • Use references/openai-harness-practices.md for the full practice-to-artifact mapping.
  • Use references/rollout-checklist.md for phased adoption in active repos.
  • Use references/wizard-cli.md for Typer wizard command flows.
  • Use assets/templates/ when creating or updating harness files.

Inputs

  • Target repository path.
  • Existing command surface (make, npm, cargo, pytest, etc.).
  • Existing CI workflows and branch protections.

Workflow

1. Baseline the repo and detect existing workflows. 2. Bootstrap harness artifacts and templates. 3. Apply all nine Harness Engineering practices. 4. Run harness audit checks and repair gaps. 5. Iterate after real agent runs.

Step 1: Baseline The Repo

  • Identify language/toolchain and canonical entrypoints.
  • Inventory existing checks, scripts, and CI jobs.
  • Record current pain points for agent runs: setup drift, unclear docs, flaky tests, missing trace IDs, slow loops.

Use a short baseline note inside PLANS.md so decisions remain durable.

Step 2: Bootstrap Harness Artifacts

Preferred entrypoint:

python3 scripts/harness_wizard.py init <repo-path> --profile control

Profiles:

  • baseline: only core harness artifacts.
  • control: baseline + control-system primitives.
  • full: control + entropy controls (nightly audit + entropy checks).

Direct shell fallback:

Run:

./scripts/bootstrap_harness.sh <repo-path>

This script installs safe defaults from assets/templates/:

  • AGENTS.md
  • PLANS.md
  • docs/ARCHITECTURE.md
  • docs/OBSERVABILITY.md
  • Makefile.harness (+ -include Makefile.harness in Makefile)
  • scripts/audit_harness.sh
  • scripts/harness/{smoke,test,lint,typecheck}.sh
  • .github/workflows/harness.yml

By default, existing files are not overwritten. Pass --force to replace template-managed files.

Step 3: Apply The Nine Practices

Implement each practice directly in repo artifacts.

1. Make Easy To Do Hard Thing

  • Ensure hard, high-value tasks are one command away (make smoke, make check, make ci).
  • Keep setup and cleanup scripted.
  • Make smoke checks cheap enough for frequent use.

2. Communicate Actionable Constraints With Compact Docs

  • Keep AGENTS.md short, concrete, and command-first.
  • Document non-obvious constraints and guardrails.
  • Keep docs close to code and update with behavior changes.

3. Structure Codebase With Strict Boundaries And Flow

  • Define module boundaries in docs/ARCHITECTURE.md.
  • Parse and validate data at boundaries; use typed contracts for internal flow.
  • Prefer one abstraction per module and one clear ownership path.

4. Build Observability In From Day 1

  • Emit structured logs/events with correlation IDs.
  • Capture key transitions in long-running workflows.
  • Define minimum observable fields in docs/OBSERVABILITY.md.

5. Optimize For Agent Flow, Not Human Flow

  • Treat context as a first-class system dependency.
  • Use PLANS.md for multi-step/multi-hour tasks.
  • Front-load durable context (scope, constraints, checkpoints) so restarts stay cheap.

6. Bring Your Own Harness

  • Standardize repo-local wrappers (Makefile.harness, scripts/harness/).
  • Wrap local infra actions in deterministic scripts.
  • Make agent behavior reproducible across machines and runs.

7. Prototype In Natural Language First

  • Draft logic and tests in prose before coding.
  • Review edge cases in prose and lock acceptance criteria.
  • Translate approved prose into code and tests.

8. Invest In Static Analysis And Linting

  • Pin formatter/linter/typechecker versions where practical.
  • Enforce checks in both local workflow and CI.
  • Run static checks before long tests to shorten failure loops.

9. Manage Entropy

  • Add periodic audits for docs drift, flaky checks, and dead scripts.
  • Keep templates synchronized with real workflows.
  • Remove stale abstractions quickly to keep agent context clean.

For a detailed artifact matrix, load references/openai-harness-practices.md.

Step 4: Validate

Run:

python3 scripts/harness_wizard.py audit <repo-path>

Treat any MISSING or FAIL result as blocking before calling harness setup complete.

Step 5: Iterate On Real Runs

  • Observe one full agent run from clean checkout to merged change.
  • Patch harness gaps immediately.
  • Re-run audit.
  • Keep AGENTS.md, PLANS.md, and architecture docs aligned with current behavior.

Adaptation Rules

  • Preserve existing project conventions and replace templates incrementally.
  • Do not overwrite user-authored files without explicit approval.
  • Keep command names stable; change internals behind wrappers.
  • Favor deterministic, scriptable workflows over ad-hoc interactive steps.

Related skills

FAQ

What does bootstrapping install?

AGENTS.md, PLANS.md, docs/ARCHITECTURE.md, docs/OBSERVABILITY.md, Makefile.harness, harness scripts and a .github/workflows/harness.yml CI workflow.

Does it overwrite existing files?

No. By default existing files are not overwritten; you pass --force to replace template-managed files.

How is setup validated?

Running the audit command; any MISSING or FAIL result is blocking before harness setup is considered complete.

DevOps & CI/CDdevopsdocs

This week in AI coding

Five minutes, every Monday - the tools, releases and tactics for developers.

unsubscribe anytime.