AI Agent Frameworks
Software development skill, available on Zeplik
AI Agent Frameworks is a ready-to-run software development skill on Zeplik. Not for authoring a Zeplik skill (use skill-creator). Ask in plain language and Zeplik applies the skill's method for you inside the conversation, on whichever AI model you prefer.
The AI Agent Frameworks skill loads automatically when your request matches it, or you can invoke it directly by typing /ai-agent-frameworks in any chat. It works with attachments, connectors, and any model that supports the task, so you get the same expert method every time without setting anything up.
What the AI Agent Frameworks skill can do
- Design LangGraph and LangChain agent architectures with state graphs and memory
- Size and compose multi-agent teams with preset configurations and heuristics
- Decompose tasks into dependency graphs and coordinate parallel agent work
- Define message protocols, file ownership, and merge strategies between agents
Try these prompts on Zeplik
Pick a prompt to open it in the Zeplik app. If you are not signed in yet, your prompt is waiting for you the moment you do.
How the AI Agent Frameworks skill works
/ai-agent-frameworks
Umbrella skill for building multi-agent systems and agent orchestration: LLM application architecture with LangChain/LangGraph, composing and sizing agent teams, decomposing tasks and coordinating parallel work, communication protocols between agents, context-driven development artifacts, track and workflow management for orchestrated pipelines, and quality guards for AI-generated work. The body of this skill is a dispatch table; the depth lives in references/. Pick the reference(s) matching the user's intent, read them, then answer. This lane covers designing and running agent systems in any framework or harness -- it is not the lane for authoring a Zeplik skill itself (that is skill-creator).
Dispatch table
| Intent | Reference |
|---|---|
| LangChain 1.x / LangGraph application architecture: agents, memory, tools, chains, state graphs | references/langchain-architecture.md (+ --details) |
| How many agents, which types, preset team configurations, sizing heuristics | references/team-composition-patterns.md (+ --preset-teams, --agent-type-selection) |
| Breaking work into tasks, dependency graphs, workload balancing, task descriptions for agents | references/task-coordination-strategies.md (+ --dependency-graphs, --task-decomposition) |
| Message protocols between agents: message types, plan approval, shutdown, anti-patterns | references/team-communication-protocols.md (+ --messaging-patterns) |
| Parallel feature development: file ownership, conflict avoidance, integration/merge strategies | references/parallel-feature-development.md (+ --file-ownership, --merge-strategies) |
| Parallel code review across quality dimensions, dedup, severity calibration, consolidated reports | references/multi-reviewer-patterns.md (+ --review-dimensions) |
| Competing-hypothesis debugging with parallel investigators and root-cause arbitration | references/parallel-debugging.md (+ --hypothesis-testing) |
| Project context artifacts agents read (product.md, tech-stack.md, workflow.md, tracks.md), scaffolding, sync | references/context-driven-development.md (+ --details, --artifact-templates) |
| Tracks as work units: spec.md, plan.md, track lifecycle for features/bugs/refactors | references/track-management.md (+ --details) |
| TDD workflow for orchestrated tasks: phase checkpoints, commits per task, verification protocol | references/workflow-patterns.md (+ --details) |
| Auditing AI-generated code for hidden debt: thin error handling, orphaned resources, hallucinated packages | references/ai-debt-detector.md |
| Long-session drift, context compaction amnesia, behavioral self-enforcement for agents | references/session-guard.md |
| Precise UI edits from visual context: element selections, annotations, screenshots | references/visual-edit-precision.md |
How to work
- Identify the intent and read the matching reference file(s) before answering; load the
--companion files when the main reference points to them. Orchestration questions usually chain references -- team sizing feeds task decomposition feeds communication protocol. - Ask ONE clarifying question when the harness or framework is ambiguous: LangChain/LangGraph vs CrewAI vs a bespoke orchestrator, and whether agents share a workspace or run isolated.
- Produce concrete artifacts: architecture diagrams in text, LangGraph/LangChain code, task decomposition tables with dependency edges and file-ownership maps, message schemas, and context artifact templates the user can drop into their repo.
- Apply the quality guards proactively: when reviewing agent output, run the ai-debt-detector checklist; when designing long-running agents, bake in session-guard practices.
- Redirect out-of-lane requests: writing a Zeplik skill goes to skill-creator; general LLM app concerns like RAG or evals go to rag-implementation, llm-engineering, or llm-evaluation; dispatching parallel subagents inside THIS harness has dispatching-parallel-agents and subagent-driven-development as sibling lanes.
Usage
/ai-agent-frameworks $ARGUMENTS
How to use the AI Agent Frameworks skill
Sign in to Zeplik
Create a free Zeplik account or sign in. New accounts start with free credits, so you can try the AI Agent Frameworks skill right away.
Describe your software development task
Ask in plain language, or type /ai-agent-frameworks to invoke the skill directly. Zeplik recognizes the AI Agent Frameworks skill and applies its method.
Review and refine the result
Zeplik returns a clear, structured answer. Ask follow-ups in the same chat to refine it or take the next step.
Source and credit
- Author
- anthropic
- License
- MIT
Adapted from the open-source wshobson/agents project and tuned to run natively on Zeplik. View source on GitHub.
Frequently asked questions
- What is the AI Agent Frameworks skill?
- AI Agent Frameworks is a ready-to-run software development skill on Zeplik. Not for authoring a Zeplik skill (use skill-creator). Ask in plain language and Zeplik applies the skill's method for you inside the conversation, on whichever AI model you prefer.
- How do I use AI Agent Frameworks on Zeplik?
- Sign in to Zeplik and ask in plain language, or type /ai-agent-frameworks in any chat to invoke it directly. The skill applies its method and returns a result you can refine in the same conversation.
- Which AI model does the AI Agent Frameworks skill use?
- Any model you choose. Zeplik works across every model in one chat, so the AI Agent Frameworks skill runs on your preferred model for the task.
- Where does the AI Agent Frameworks skill come from?
- The AI Agent Frameworks skill is adapted from the open-source wshobson/agents project (MIT) and tuned to run natively on Zeplik. The original source is linked on this page.
- How much does the AI Agent Frameworks skill cost?
- Using the skill is free to start. You only spend Zeplik credits when the assistant runs, and new accounts begin with free credits.
Related software development skills
- .NET BackendBuild ASP.NET Core 8+ backends with EF Core: auth, background jobs, production API patterns
- Advanced Git WorkflowsUse for advanced Git surgery: interactive rebase, cherry-pick, bisect, reflog recovery, and history cleanup before merging. Not for parallel worktree workflows (use using-git-worktrees).
- Adversarial Code ReviewHunt for bugs in code the user shares by assuming defects exist and attacking the code through several distinct lenses, then report severity-ranked findings with evidence. Use for "review this", "what could go wrong", "bug hunt", or pre-merge scrutiny of a change. Read-only, it reports problems and does not rewrite the code. Not for style cleanup (use simplify-code) or for writing new code.
- Algolia SearchAdd Algolia search: indexing strategies, React InstantSearch, relevance tuning, search-as-you-type
- Android CI/CDAutomate Android CI/CD to Google Play: keystore, GitHub Secrets, multi-stage release workflow for RN, Flutter, native
- Angular (v20+)Modern Angular v20+: Signals, standalone components, zoneless apps, SSR/hydration, reactive patterns
More on Zeplik
Try AI Agent Frameworks on Zeplik
Every model, one chat. Bring the AI Agent Frameworks skill into your next conversation and let the assistant do the work.