clawd-conduit
Plugin / AgentWire all 15 Claude Code hook events into the Clawd on Desk desktop pet — and turn permission requests into a blocking desktop card.
- hooks
- desktop
- notifications
- claude-code
Pydantic AI
Type-safe agent framework from the Pydantic team — structured outputs and typed tools for Claude.
Pydantic AI applies the Pydantic ethos to agents: typed tool signatures, validated structured outputs, dependency injection, and streaming, with first-class Anthropic model support. The default choice for Python teams that already trust Pydantic for data validation and want the same guarantees in their agent layer.
What it does
- Agent framework from the Pydantic team — type-safe, structured outputs validated by Pydantic models
- Model-agnostic: OpenAI, Anthropic, Gemini, and more behind one interface
- Optional extras like pydantic-ai[temporal] for durable execution and pydantic-ai-harness
- Installed via uv add (or pip) like a normal Python package
When to use it
For Python developers who already think in Pydantic models and want agents whose outputs are guaranteed to match a schema — production services, data pipelines, typed tool calls.
Getting started
uv add pydantic-ai Optional extras: uv add "pydantic-ai[temporal]", uv add pydantic-ai-harness.
Things to know
It's a framework, not a turnkey agent — you write the agent loop, tools and prompts in Python.
Repository
pydantic/pydantic-aiCharted
Related on the map
CometixCode
Plugin / AgentAn unofficial Rust 1:1 reimplementation of the Claude Code terminal UI, ported file-for-file onto an iocraft TUI framework.
- terminal-ui
- rust
- reimplementation
- unofficial
Abide
Plugin / AgentHook-level enforcement for the natural-language rules in AGENTS.md and CLAUDE.md — every edit is judged and violations are repaired in the same turn.
- hooks
- code-review
- rules
- agents-md