Set up the codebase harness
SkillAI & modelsThis skill sets up a repository so an AI agent can work in it reliably. Once it is in place, the agent gets clear documentation, a development stack it can start with one command, and checks that verify its work before anything ships. It is designed for onboarding an agent to a new or unfamiliar codebase.
Available today. Use it from your connected AI after setup.
No other account needed.
Add the skill and run it on the repository you want to onboard, ideally one that is new or unfamiliar to your agent. Once setup is complete, the agent can start working the codebase with documentation, a runnable stack, and verification already in place.
Then ask your AI: use the Set up the codebase harness skill
What your AI can do with it
- Create map-style documentation that shows how the codebase fits together
- Add custom lints that enforce the project's own rules
- Set up a development stack the agent can run with a single command
- Add an end-to-end gate that checks work before it ships
- Run a verify-before-ship loop so changes are confirmed working
- Keep commits clean and control entropy so the codebase stays manageable over time
What this skill tells your AI
The instructions your AI receives, as published by ai-builder-club/skills in skills/setup-codebase-harness/SKILL.md and read by ahel’s review.
Harness engineering: the model is fixed — what you engineer is the scaffolding around it (the environment, the docs, the feedback loops) so an agent can build and verify software with minimal human attention. Humans steer; agents execute. Your job is to make the repo legible, executable, and verifiable.
Work incrementally and depth-first: assess what exists, build the one missing capability, use it to unlock the next. Don't boil the ocean — set up what the repo actually needs. When the agent struggles, the fix is almost never "try harder" — ask "what capability is missing, and how do I make it legible and enforceable?" and add it.
This skill orchestrates the focused sub-skills: dev-local-setup,
e2e-setup, crabbox-setup (cloud/parallel), and verifier-setup
(scaffolds a repo-specific /verify loop; supersedes the older pr skill).
0. Assess
Survey the repo: stack, package manager, services/ports, infra deps, existing docs/tests/CI, and the implicit rules (buried in READMEs, PR comments, people's heads). Note what's missing per pillar below.
1. Legible — the agent can reason about the repo
What the agent can't see doesn't exist. Knowledge in chat threads / heads is invisible — push it into versioned, repo-local artifacts.
- a) Map, not manual. Shrink the root agent doc (
AGENTS.md/CLAUDE.md) to a ~100-line table of contents: one-line overview, project tree, golden rules (the hard invariants), and a "where to look" table. Move the depth into a structureddocs/system-of-record (architecture, frontend, testing, domain topics) with adocs/index.md. A monolithic instruction file rots and crowds out the task — keep the map small and stable, disclose detail progressively. - b) Custom lints with remediation. Promote the prose golden rules into
mechanical checks — human taste captured once, enforced everywhere, every run.
One lint per invariant (layering / dependency direction, naming, no-
any, forbidden imports, file-size, structured logging). Write the error message to inject the fix ("X isn't allowed here — do Y") so the remediation lands in agent context. Wire them into the repo's linter + CI. - c) Queryable code graph. Index the repo with
codebase-memory-mcpso the agent traces callers, data flow, and architecture from a knowledge graph instead of blind grepping — faster, more precise navigation on large codebases. - d) (later) Keep docs honest. A freshness / doc-gardening pass that flags docs that no longer match the code and opens fix-up PRs.
2. Executable — the agent can run & drive the app
dev-local-setup→ a one-command, reproducible local stack (scripts/dev-local.sh up) running every service + infra.- Make the app drivable: browser via the
playwright-cliskill (installed byverifier-setup); logs reachable. crabbox-setup→ an isolated cloud box per agent — the parallel-safe counterpart to dev-local. Reach for it when loops run concurrently: one laptop can't host N full stacks (fixed ports, one Docker daemon, one DB), and per-worktree local doesn't fix it — the worktrees still share the host. crabbox gives each agent its own stack + an in-box browser, so parallel verification never collides.- Advanced: a local, ephemeral observability stack (queryable logs/metrics) for perf/reliability prompts.
3. Verifiable — the agent can prove it works
e2e-setup→ a trustworthy e2e gate: real flows (not bypass), a reusable auth/session helper, layered client → server → product assertions, video/trace evidence, sandbox-only external services.verifier-setup→ scaffolds a repo-specific/verifyskill (the verify-before-ship loop): a fresh verifier sub-agent drives the real app to confirm the just-built feature works; the main agent fixes until green, runs the codified regression sweep, and opens a PR with a reviewable proof link. Add the session helper so the verifier can reach login-gated features.
4. Others — keep it coherent over time
- Commit hygiene: conventional commits + format/lint on commit (e.g. husky lint-staged + commitlint). Keep merge gates light — at high agent throughput, corrections are cheap and waiting is expensive.
- Garbage collection: encode "golden principles", then run periodic cleanup passes that open small refactor PRs — pay tech debt down continuously, not in painful bursts. Human taste captured once, enforced on every line.
- Agent-to-agent review for correctness-critical changes (independent reviewers, not self-review).
Order & what you leave behind
1a (map) → 2 (dev-local) → 3 (e2e + verifier-setup), then 1b (lints) and 4 as the
repo matures. The artifacts — slim map + docs/, scripts/dev-local.sh, an e2e/
suite, the generated /verify skill, and custom lints — are each a reusable, legible capability
that compounds. Prefer "boring", composable, stable tech the agent can fully model.
Signals
- GitHub stars
- 1k
- Forks
- 157
- Last commit
- Jul 2026
Advanced
- Catalog kind
- skill
- Gateway key
setup-codebase-harness- Source
- github.com/ai-builder-club/skills