By JasGujral
Spec-driven TDD framework for Claude Code — Socratic brainstorm, provable completeness gate, codebase knowledge graph, persistent self-healing memory, token governance, and human-gated writes.
mokata · Socratic pre-spec exploration — explore approaches WITH the user, one question at a time, and HARD-GATE the spec behind explicit approval. Runs standalone or as the front of mokata's pipeline.
mokata · The command palette — every /mokata command and bundled skill on one screen, with gate markers. Read-only.
mokata · Start from a reproducer and a failing test, then fix.
mokata · Plan a manual chain of skills to run in order — each step keeps its own gate.
mokata · Reproduce first, capture in a failing test, then fix.
mokata · API & interface design — contract-first, blast-radius on every change.
mokata · Socratic pre-spec exploration — explore approaches WITH the user, one question at a time, and HARD-GATE the spec behind explicit approval. Runs standalone or as the front of mokata's pipeline.
mokata · Browser testing (DevTools) — live runtime captured via the existing MCP surface, fed to the test gate.
mokata · Start from a reproducer and a failing test, then fix.
mokata · CI/CD & automation — the quality-gate pipeline + Scorecard posture, not a suggestion.
Executes bash commands
Hook triggers when Bash tool is used
Modifies files
Hook triggers on file write and edit operations
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
The memory + seatbelt for your AI coding agent. Spec-driven TDD for Claude Code — knowledge-aware, self-healing memory, human-gated, local-first. It remembers your project and stops the agent shipping the wrong thing.
📖 Full documentation: https://mokata.ai/ — quickstart, tutorials, concepts, and the complete CLI + plugin reference.
An AI agent, mid-task, tries two bad changes — ship code with no spec, and stash an AWS key while
waving its own write through. mokata stops both. (Real output; run it yourself with
mokata init --profile standard --yes in a scratch repo.)
$ mokata run develop # the agent tries to write code with no spec...
[BLOCKED] spec-persisted — no saved spec — draft and emit it first (/mokata:spec); the completeness gate must pass before implementation.
$ # ...then tries to stash a secret, approving its own write with approve=True:
status: proposed committed: False
`approve`/`confirm` no longer commit: an approval is MINTED BY A HUMAN out-of-band (`mokata approve <id>`) and can only be REFERENCED here by id. Typing approve=true is not consent — it never was.
$ mokata audit # the gate decision is on the ledger
audit ledger — 1 entry:
#1 gate gate=spec-persisted phase=develop decision=blocked reason=no saved spec ...
A model cannot approve its own writes at all. It can only propose one and hand you an id; the approval is minted by you, out-of-band, in a process the model isn't driving — and it is single-use, content-hashed, and expires. And if you do approve that write, the secret-guard still refuses it: approval is a methodology gate, never a security override.
There is a third block the agent can't route around either — reaching past mokata's tools for the
editor's own Write trips a PreToolUse hook (exit 2) and the file is never touched. It's in the
full demo.
The full copy-paste demo + a 60-second screencast script: mokata catches a bad change. See every differentiator run in differentiators in action.
A quick tour of the workflow and skills — brainstorm → spec → test → develop → review → ship, with governance, memory, and the always-on status badge.
https://github.com/user-attachments/assets/a6940119-1edb-4017-b935-cd48b80b4ea9
▶ Video not playing inline? Watch it here.
npx claudepluginhub jasgujral/mokata-ossSPEC-First development workflow with TDD, Ralph Loop, and autonomous agent coordination for Claude Code
Ryan's riff on obra/superpowers: strong guidance for Claude Code as a software development assistant
Test-driven development methodology with red-green-refactor cycles and code review
Use Claude Code like a Director, not a Programmer. Complete toolkit with 27 commands, 14 agents, 32 skills, and TDD-based Auto-Loop.
Describe your goal, approve the spec, then step away — Claude and Codex loop together until it's right.
Helder's personal SDLC toolbelt for AI coding agents — from spec to ship. Bundles the tracer-bullet workflow alongside TDD, code review, audits, and shipping skills.