Self-learning QA plugin — builds testing knowledge, verifies BDDs, guards regressions
Run superpowers brainstorming for a feature, then automatically generate BDD scenarios from the resulting spec
Exploratory testing to build knowledge about the system — discover behaviors, map flows, and grow the QA knowledge graph
Generate implementation-agnostic BDD scenarios from a spec or plan document and drop them in the QA dropbox (qa/bdds/)
Set up a new claudity-assurance QA environment — interview, scaffold directory, configure interaction surface
Reset the QA environment — full re-onboard with archival, or troubleshoot and adapt the documentation approach
This skill should be used when a user wants to set up a new claudity-assurance QA environment for a project. Trigger phrases include "onboard this project", "set up QA", "initialize QA environment", "start QA for this project", "c-a onboard", "new QA project", and "I want to start testing this system". Also activates when a user starts a session in an empty directory and asks about testing or verification.
This skill should be used when a QA session needs to verify BDD scenarios against a running system. Trigger phrases include "verify these BDDs", "test the new features", "run verification", "check the bdds", "verify against the system", "start QA verification", "what needs testing", and "run the BDD scenarios". Also activates when new .feature.md files appear in the bdds/ directory and the user asks to test them.
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.
A self-learning QA plugin for Claude Code that gives solo developers an independent verification layer for their work.
It operates as a black-box tester — it never sees your source code. Instead, it builds its own knowledge about your system through interviews, exploration, and testing, then verifies behavioral specifications (BDDs) against the running system. Over time, it accumulates institutional QA knowledge that gets better with every session.
When you're a solo developer, you're writing code and testing it yourself. That means your tests are biased by your knowledge of the implementation — you test what you built, not what the system should do. Bugs hide in the gap between those two things.
Claudity-assurance fills that gap by creating a QA session that:
Claudity-assurance operates across two Claude Code sessions:
Session 1 (Dev) Session 2 (QA)
───────────────── ────────────────
Your normal dev work Isolated QA directory
│ │
├─ Brainstorm features ├─ Onboard: interview about the system
├─ Write specs & plans ├─ Build interaction surface (browser, API, etc.)
├─ Implement code ├─ Verify BDDs against running system
├─ Generate BDDs ──── qa/bdds/ ─────────>├─ Produce verification reports
│ ├─ Run regression checks
│ ├─ Update knowledge graph
│ ◄── results/ ────────────├─ Evidence (screenshots, logs, notes)
└─ Fix failures └─ Get smarter every session
Session 1 (Dev) — Your normal development session. After brainstorming or implementing a feature, generate BDD scenarios that describe what the system should do in plain behavioral terms. These get dropped into a shared qa/bdds/ directory.
Session 2 (QA) — A separate Claude session running in an isolated QA directory. It picks up BDD scenarios, tests them against the running system as a black box, and builds a growing knowledge base about how to interact with and verify the system.
The two sessions are deliberately isolated:
qa/bdds/ — it cannot read QA results or docsqa/ — it cannot read source code, specs, or implementation filesqa/bdds/) is the only shared artifactThis separation ensures the QA session tests behavior, not implementation.
/plugin marketplace add https://github.com/look-itsaxiom/claudity-assurance.git
/plugin install claudity-assurance
Or for local development:
claude --plugin-dir /path/to/claudity-assurance
You have a project and want to add QA verification.
1. Create the QA environment
cd your-project
mkdir qa
cd qa
claude
/claudity-assurance:onboard
Claude will interview you about the system — what it is, how to access it, what tools to use for testing. It then scaffolds the QA directory and writes its initial knowledge docs.
2. Generate BDDs from your dev session
In a separate Claude session in your project root:
/claudity-assurance:generate-bdds
This reads your most recent spec or plan and produces behavioral scenarios in Gherkin format, dropping them in qa/bdds/.
3. Verify in the QA session
Back in the QA session:
/claudity-assurance:verify
Claude picks up the BDDs, tests each scenario against the running system, collects evidence, checks for regressions, and produces a verification report.
Starting fresh with no code yet.
1. Create the QA directory and onboard
mkdir my-project/qa
cd my-project/qa
claude
/claudity-assurance:onboard
During onboarding, Claude will help you define initial BDD scenarios and test them — building its knowledge base from scratch.
2. Brainstorm with BDDs
In your dev session, use the combined workflow:
/claudity-assurance:brainstorm-with-bdds
This runs the full brainstorming process and automatically generates BDDs from the approved spec.
3. Iterate
Implement features in Session 1, generate BDDs, verify in Session 2. The QA session gets smarter each cycle.
| Command | Purpose |
|---|---|
/c-a:generate-bdds | Generate BDD scenarios from a spec or plan document |
/c-a:brainstorm-with-bdds | Brainstorm a feature, then generate BDDs from the result |
npx claudepluginhub look-itsaxiom/claudity-assurance --plugin claudity-assurance10 coordinated QA agents for Claude Code: chains diff analysis, AC compliance, test scenarios, browser validation, bug reports, and automation code generation.
QA skills for automation domain.
Comprehensive Behavior-Driven Development principles, practices, and collaboration patterns.
AI test generation with Ralph-loop quality gate: coder → reviewer → iterate
Comprehensive skill pack with 66 specialized skills for full-stack developers: 12 language experts (Python, TypeScript, Go, Rust, C++, Swift, Kotlin, C#, PHP, Java, SQL, JavaScript), 10 backend frameworks, 6 frontend/mobile, plus infrastructure, DevOps, security, and testing. Features progressive disclosure architecture for 50% faster loading.
Develop, test, build, and deploy Godot 4.x games with Claude Code. Includes GdUnit4 testing, web/desktop exports, CI/CD pipelines, and deployment to Vercel/GitHub Pages/itch.io.