Where agentic tools are forged. Claude Code plugins for AI engineering.
Context is everything. Before agents can act, they must understand.
This foundry grew from years of work on context engineering - the practice of giving AI systems the right information at the right time. It started with Codebase Context and the .context.md specification, which laid the pattern for the AGENTS.md standard now adopted across the industry. The same principles drive MCP, skills, and every plugin here.
Our principles:
- Context over prompting - Rich, structured context beats clever prompt tricks
- Open standards - Portable formats that work across tools and vendors
- Consulting-grade - Production patterns from real client engagements
- Evidence over claims - Test it, measure it, prove it works
These plugins encode repeatable solutions from AI engineering consulting. Each represents a pattern that worked in production, extracted and packaged for reuse.
Full story: codebasecontext.org
/plugin marketplace add agentic-insights/foundry| Plugin | Description | Install |
|---|---|---|
| aws-agentcore-langgraph | Deploy LangGraph agents on AWS Bedrock AgentCore | /plugin install aws-agentcore-langgraph@agentic-insights |
| baml | Type-safe LLM extraction with BAML codegen | /plugin install baml@agentic-insights |
| build-agent-skills | Build skills with Agent Skills open standard | /plugin install build-agent-skills@agentic-insights |
| copywriter | UX/marketing copy + Chain-of-Density summarization | /plugin install copywriter@agentic-insights |
| adversarial-coach | Adversarial code review (Block g3 research) | /plugin install adversarial-coach@agentic-insights |
| vhs-recorder | Record terminal demos with Charm VHS | /plugin install vhs-recorder@agentic-insights |
| para-pkm | PARA knowledge management system | /plugin install para-pkm@agentic-insights |
aws-agentcore-langgraph
Deploy LangGraph 1.0 agents on AWS Bedrock AgentCore.
- Skill:
aws-agentcore-langgraph- Runtime, memory, and Gateway integration - CLI:
agentcore configure|launch|invoke|destroy
baml
Type-safe LLM extraction with BAML. MCP-powered codegen with LangGraph integration.
- Skill:
baml-codegen- Generate code, schemas, tests from requirements - Agents:
baml-architect,baml-debugger,baml-tester - Commands:
/baml-init,/baml-schema,/baml-test,/baml-diagnose - Requires: MCP server
baml_Docs(optional:baml_Examples)
Documentation → | Credits: Fry
build-agent-skills
Build portable skills with the Agent Skills open standard.
- Skill:
working-with-agent-skills- Create, validate, publish skills - CLI:
skills-ref validate|inspect
copywriter
Expert copywriting for UX, marketing, and product content with Chain-of-Density summarization.
- Skills:
copywriter(UX copy, landing pages, emails),chain-of-density(iterative compression) - Agent:
cod-iteration- Single-iteration density worker
Documentation → | Credits: daffy0208/ai-dev-standards
adversarial-coach
Adversarial code review based on Block's g3 dialectical autocoding research.
- Command:
/coach [requirements-file]- Independent implementation review - Agent:
coach- Fresh context reviewer, catches missing auth/error handling
vhs-recorder
Record terminal sessions with Charm's VHS for CLI demos and tutorials.
- Skill:
vhs-recorder- Create VHS tape files with timing control - Examples:
basic-demo.tape,cli-tool-demo.tape,git-workflow.tape
para-pkm
PARA (Projects, Areas, Resources, Archives) knowledge management system.
- Skill:
para-pkm- Organize knowledge bases with PARA methodology - Scripts:
init_para_kb.py,archive_project.py,validate_para.py
| Plugin | License |
|---|---|
| aws-agentcore-langgraph | Apache-2.0 |
| baml | Apache-2.0 |
| build-agent-skills | Apache-2.0 |
| copywriter | Apache-2.0 |
| adversarial-coach | Apache-2.0 |
| vhs-recorder | Apache-2.0 |
| para-pkm | Apache-2.0 |
GitHub Issues for bugs and feature requests.
