Independent authorization plane for AI agents — approve, escalate, or deny spend, tools, credentials, and provisioning. Wires the hosted wallet MCP; the host must ask before acting.
On Claude Fable 5.1 (or Opus), Claude's helpers (subagents) inherit your expensive model and can start helpers of their own — one web research pass can burn a session's usage limit in minutes, and an instruction to be frugal doesn't survive context compaction. frugal-subagents builds the limits in: helpers run on a…
★not rated 3 6d agoA
tokens not measured
originalMIT
Memory that survives compaction: your agent's decisions with their why, permanent dead-ends, and causal chains — built automatically from the Claude Code transcripts you already have. Bundles the NodeDex MCP connection + the memory-discipline skill.
★not rated 2 2mo agoA
tokens not measured
AGPL-3.0
Multi-model LLM Council + parallel AI window for Claude Code. Run structured council reviews across Gemini, GPT, DeepSeek and more — or fork a conversation to any model and fold the results back.
★not rated 2
changed 3d agoA
tokens not measured
originalMIT
Preflight risk linting for MCP & agent tools. Before you add or edit an agent-tool config, CallLint recommends scanning the blast radius — SAFE / REVIEW / BLOCK / UNKNOWN with evidence. Advisory and non-blocking; never executes the server it judges.
★not rated 2 todayA
tokens not measured
originalApache-2.0
Give your agent a persistent identity on AIIM: a live network where AI agents chat, make friends, found projects, and earn reputation that compounds across sessions.
★not rated 2 todayA
tokens not measured
originalMIT
Install and connect Firekeep — one self-hosted Keep that carries knowledge, working context, coordination, and evidence across AI agents — then verify it with the product's own doctor.
Codexa for Claude Code. MCP server exposes evidence-backed repo context tools; SessionStart injects repo context, PreToolUse saves an implicit pre-edit baseline, Stop runs the post-edit drift review and surfaces blocking drift to the model. Slash commands wrap the codexa CLI.
pi-coding-agent running in a container — interactive shell, one-shot exec, an HTTP REST API, an OpenAI-compatible chat endpoint, an MCP server, a Telegram bot, and a cron scheduler, all on one image.
Skills that teach any agent to use chat-recall: recall a decision and why it was made, find credentials pasted into past sessions, rank what is worth fixing in a repository, and track it on a shared board. Bundles the MCP server so install wires everything in one step.
Real-time collaboration between two Claude Code sessions over a shared IRC channel: a built-in chat hub, self-naming peers, and automatic message delivery via hooks.
★not rated 2 2mo agoA
tokens not measured
originalMIT