Claude Code plugin marketplace for building agentic systems. Four plugins covering LangGraph development, development practices, BDD quality assurance, and the neograph graph compiler.
# Add the marketplace
/plugin marketplace add postindustria-tech/agentic-toolkit
# Install specific plugins
/plugin install langgraph-dev@agentic-toolkit
/plugin install dev-practices@agentic-toolkit
/plugin install qa-bdd@agentic-toolkit
/plugin install neograph-dev@agentic-toolkitUpgrading from pi-dev? See UPGRADE.md for instructions.
Tutorial-grounded patterns and best practices for LangGraph development.
| Component | Count | Status |
|---|---|---|
| Skills | 21 | Production ready |
| Commands | 5 | WIP |
| Agents | 1 | WIP |
Skills cover: state management, conditional routing, conversation memory, ReAct agents, multi-agent supervisors, human-in-the-loop, streaming, structured output, subgraphs and composition, basic and corrective RAG, document processing, error recovery, parallel execution, performance optimization, prompt engineering, memory store and knowledge, deployment patterns, testing agentic systems.
Commands (WIP): create-workflow, create-state, create-tests, create-agent, create-deployment.
Agent (WIP): workflow-validator.
Quality-driven development workflows with beads task management integration.
| Component | Count | Status |
|---|---|---|
| Skills | 8 | Production ready |
| Agents | 5 | Production ready |
Skills:
| Skill | Purpose |
|---|---|
execute |
Molecular task execution — atoms chained in sequence, survives context compaction |
tdd-workflow |
Test-driven development cycle with sacred rule enforcement |
session-completion |
End-of-session checklist ensuring no work is lost |
code-review |
Multi-agent code review across 8 dimensions |
obligation-test |
Per-obligation behavioral test derivation |
remediate |
Fill entity test stubs batch-by-batch with TDD |
test-audit |
Audit test sources of truth |
reclassify |
Reclassify and reorganize tasks |
Review agents: review-consistency, review-dry, review-layering, review-python-practices, review-testing.
Quality assurance tools for pytest-bdd projects and Claude Code plugin skills.
| Component | Count | Status |
|---|---|---|
| Skills | 4 | Production ready |
Skills:
| Skill | Purpose |
|---|---|
skill-qa-audit |
Read-only audit of plugin skills against official documentation |
skill-perfection |
Audit and fix skills in a single pass with optional Python preflight |
inspect-steps |
Context-aware BDD step assertion inspection |
step-development |
BDD step development workflows with AST-based anti-pattern detection |
Development toolkit for the neograph declarative LLM graph compiler.
| Component | Count | Status |
|---|---|---|
| Skills | 5 | Production ready |
Skills:
| Skill | Purpose |
|---|---|
neograph-dev-codebase-nav |
Import map, layer discipline, deferred import budget, module ownership |
neograph-dev-bug-fix |
TDD bug fix workflow with mutation verification and past bug catalog |
neograph-dev-test-design |
Test file layout, fake infrastructure, three-surface parity, Hypothesis |
neograph-dev-rendering |
Two prompt systems, rendering dispatch hierarchy, BAML parity invariant |
neograph-dev-lint |
Lint architecture, extension guide, CLI wiring checklist, template resolver |
Add to your project's .claude/settings.json:
{
"extraKnownMarketplaces": {
"agentic-toolkit": {
"source": {
"source": "github",
"repo": "postindustria-tech/agentic-toolkit"
}
}
},
"enabledPlugins": {
"langgraph-dev@agentic-toolkit": true,
"dev-practices@agentic-toolkit": true,
"qa-bdd@agentic-toolkit": true,
"neograph-dev@agentic-toolkit": true
}
}Planned additions:
dspy-dev- DSPy patterns and optimizationagent-sdk-dev- Anthropic Agent SDK toolkit
MIT - see LICENSE