#ai-agents

Workspace context manager for AI coding agents: name your Claude Code sessions, resume them in one keypress, and carry context across sessions

Adaptive JSON-to-agent terminal output with exact token-aware format selection

🔷 Recursive decomposition workflow for AI coding agents — small confirmable design steps instead of mega-diffs. Installs /prism slash commands for Claude Code, Cursor, Codex, Gemini & more

Agent-aware tmux IDE — sidebar, session registry, and responsive panes for Claude/Codex workflows. Adapter-based; host integrations are pluggable.

The automation engine built for AI agents. Workflows, memory, and 100+ integrations — one API key.

VIP native passthrough for Anthropic & OpenAI through BlockRun — official Go SDKs, x402 USDC on Base, zero response reshaping.

Herrscher — Discord gateway plugin: adapts Discord (via dctl) onto the NATS message plane.

Herrscher — Obsidian memory plugin: a co-edited markdown knowledge graph (Project/Architecture/Production/Session/Decision/User nodes) implementing the contracts.Memory port.

Herrscher Orchestrator plugin: the default conversation-policy port — primes each turn from Memory and records a bounded rolling transcript.

Git Issues Extension for Agentic AI Task Management

A schema-first GraphQL MCP server that avoids the 1MB introspection limit via typeNames/typeDepth scoping — built for federated topologies.

Python SDK for agent spend governance around paid tool calls across MCP, LangGraph, OpenAI, Claude, and Gemini: spend authorization, evidence receipts, refunds, disputes, and settlement.

Terminal-native note system for agent-driven, scriptable workflows

Agent-native SSH multiplexing CLI

Markdown → one .slides.html deck that edits itself in the browser. Ships agent skills.

Cryptographically-authenticated, cross-machine agent-to-agent chat for Claude Code, over MCP

Your agent forgets; nem doesn't. mnemonics encodes a session's decisions, edge-cases and facts into immutable, recallable structure — versioning agent context the way git versions code, shared across Claude & Codex.

Budgeted web research and datasets for any MCP-speaking agent.

A plan-file factory for Claude Code — and the console that watches it run. Engine: parallel headless workers in git worktrees, script-owned gates, unattended drive loop (TypeScript, zero deps). Console: live web observability over driven repos.

Edit on Mac. Build and test in Linux. Local workspaces for developers and coding agents.

Personal Skill collection TUI for AI coding agents · 面向 AI 编程助手的个人 Skill 收藏终端应用

Booboo — the unified operational brain. Wiring + knowledge + memory + agents + crons fused into one rooted, queryable 3D graph. REST + MCP + million-node 60fps viewer.

Your AI remembers what it did, proves it, and gets smarter every session

该仓库未发布描述。

A stateful, graph-based agent runtime for Go — LangGraph's core ideas in 419 dependency-free lines

Turn failed build logs into cited evidence packets for Claude, Codex, CI, and humans - locally.

Official Go SDK for xShellz sandboxes — spawn a throwaway, kernel-isolated Linux box and run commands in it from your code. go get github.com/xshellz/xshellz-go

Evaluate LLMs side-by-side. Benchmark AI models and coding agents across providers like OpenAI, Google, Anthropic, DeepSeek, and more. Supports custom tasks, structured JSON responses, tool use, and LLM-as-judge validation. Originally created by Petr Malik as MindTrial.

TraceOn is an MCP server that Claude Code calls after editing code to verify the change actually works end-to-end. It runs a Playwright test the agent wrote, correlates browser evidence with backend traces and logs via OpenTelemetry, ranks the evidence into a compact packet, and returns a structured VerificationResult that the agent reasons over.

The Makefile of agent orchestration. Terminal agent coordination without the framework tax. No Python. No cloud. No magic. Just tmux, TOML, and a binary.