Skip to main content
Workflow

Plan complex work, then ship it with a team that proves every step.

Plan & Build researches a task across multiple rounds, writes a complete implementation plan, breaks it into an owner-assigned spec, then executes — a builder per task, a validator confirming each one, and completion gated on real evidence.

  • Research → plan → build, end to end
  • Every task gated on evidence
  • Runs locally with your Claude key

kr 2 700one-time · lifetime updates · install →

/ 01 The job

What runs when you install the plugin.

Five commands, one pipeline. doctor provisions the stack first; then research becomes a plan, the plan becomes a verifiable spec, and the spec gets built — nothing marked done without proof.

  1. 01pre-plan researches the task across multiple rounds and writes a complete implementation plan to plans/.
  2. 02plan-w-team turns the plan into an owner-assigned, verifiable spec in specs/. Planning only — it never builds.
  3. 03build deploys a builder per task and gates every task on verification evidence before it can be marked done.
  4. 04A validator independently confirms each task against its acceptance criteria — independent evidence, not the builder's self-report.
  5. 05doctor provisions the stack first — uv, optional Python, git — so the quality gates have what they need. The hooks fail open if uv is absent.
/ 02 What's inside

Every artefact you're getting.

A plugin is a folder. Here's everything in it — two agents, two skills, five commands, six hooks, and the plan template they all follow. The heaviest plugin in the catalogue, fully on show.

2 Agents

Domain experts with baked-in methodology, hot-loaded.

buildersonnet-4-5Completes ONE assigned task end-to-end and reports verification evidence.
validatorsonnet-4-5Read-only verification of ONE task against acceptance criteria — independent evidence, not self-report.

The whole plugin sits at ~/.claude/plugins/plan-and-build/ — every agent, every hook, and the plan template is a file you can read, edit, and version.

/ 03 Tools it reads + writes

Will it work with your stack?

Reads from
Your projectGit
Writes to
plans/specs/Your codebase
Requirements: Claude Code and git. uv is recommended (doctor installs it; the gates fail open without it) and pulls ruff and ty on demand for Python lint/type checks. No third-party API key beyond your own Claude key.

Better with Core Methodology installed — the commands invoke its disciplines (validation-first, verification-before-completion, deliverable-review) when present, and degrade gracefully when not. It does not re-bundle them.

/ 04 Install & price

Four steps. From buy to first run.

Buy on the right. On purchase, we invite your GitHub account to the private plugin repo you bought — the Tenkr marketplace itself is public — and email you the two commands below.

1Install Claude Code (Anthropic's free CLI). Already have it? Skip.
2Buy. We invite your GitHub account to the private plugin repo you paid for. The Tenkr marketplace itself is public.
3In Claude Code, run /plugin marketplace add TenkrAS/tenkr-plugin-marketplace to register the public catalogue. Exact command in your purchase email.
4Run /plugin install <plugin>@tenkr-plugin-marketplace — Claude Code clones it from your private repo and you're live. Type the plugin's command to start.
/ 05 Plugin-specific FAQ

What people ask before buying.

/ 01Do I need Core Methodology too?
No — it's a soft dependency. With Core Methodology installed, the commands invoke its disciplines; without it, they degrade gracefully. Better together, not required — and it doesn't re-bundle the methodology skills.
/ 02What if I don't have uv?
The pipeline still runs. The lint, type, and evidence hooks fail open until uv is provisioned — run doctor to install it and turn the gates back on.
/ 03Does it touch my whole repo?
It's project-agnostic. Plans go to plans/, specs to specs/, both created on first use; build tasks write wherever the spec says.
/ 04How is "evidence-gated" actually enforced?
By hooks. A SubagentStop gate blocks any task completion without proof, and PostToolUse ruff / ty gates lint and type-check Python as it's written.