Browse agent skills

Manages discovery, parsing, matching, and registration of SKILL.md skill files. Supports automatic discovery at both project and user level.

Agent working memory pipeline: LLM-driven compaction, LLM-free pruning, archive persistence, multi-source AGENTS.md memory, writable cross-session auto-memory, operational learning, and workspace file-state snapshots.

Pluggable interception and transformation engine spanning every stage of the agent lifecycle. Pure hook engine only — the permission policy system, sandbox safety assessment, and failure-signal diagnostics were split out to [`@x-otto/guard`](../guard) (RF

Promptless commands and public agent skills for integrating voice agents with Egma.

> GitHub Copilot subscription provider plugin.

Interactive coding agent (Pi) with UiPath LLM Gateway and BYO provider support.

NestJS integration for sms.ir - Dependency Injection, forRoot/forRootAsync configuration, and a zero-dependency client for the sms.ir REST API.

Coding agent CLI with read, bash, edit, write tools and session management

Codex orchestrator workflow skill: main agent delegates token-heavy work to subagents.

WhatsApp gateway CLI (Baileys web) with Pi RPC agent

Clause-driven PDF library in pure TypeScript. Every behaviour tied to an ISO 32000 clause; what cannot be tied is not claimed.

Official Rasputin AI SDK for Elysia.

Official Rasputin AI SDK for Node and Bun.

Core error capture, stack parsing, and ingest transport for Rasputin AI.

CLI tool for AI Dialogue (aichat) via AceDataCloud API

Gives text-only opencode models (DeepSeek, Qwen, GLM, ...) real image reading: pasted screenshots are auto-described via free OpenRouter vision models, with a local tesseract OCR fallback.

Cut prompt and API token cost on OpenAI and Anthropic without silently weakening requirements. Deterministic planning, no extra model call.

Bootstrap the agents-atlas convention (.agents/) into a project. Replaces `cp -r` + `ln -s`.

WorkBuddy Connector MCP runtime and Skill companion for @xqyz/xq-cli.

Local-first LLM inference SDK for JavaScript & TypeScript. Run GGUF models on your own hardware via llama.cpp, with an App/Function/Sandbox API and a 10x cost cut over always-remote inference.

Codex-shaped shell tool and prompt adapter for pi coding agent

FaceAPI: AI-powered Face Detection & Rotation Tracking, Face Description & Recognition, Age & Gender & Emotion Prediction for Browser and NodeJS using TensorFlow/JS

On-device Expo OCR for Afghan identity documents (ML Kit/Vision + PP-OCR, optional B&W preprocess)

Framework-agnostic AI client, one interface, two backends. Server: NVIDIA NIM (OpenAI-compatible, key from env NVIDIA_API_KEY, server-side only). Client: Puter.js (keyless, user-auths, :free models). Shared chat(messages, opts) contract.

n8n community node for Zavu — one API for SMS, WhatsApp, Telegram, Email, Instagram, Messenger and Voice.

Embed a Whissle voice or text agent into any web app. Create/configure the agent on whissle.ai, then run a live conversation with a publishable key.

Command-line interface for the Whissle Voice Agents platform — configure agents, run them, and pull your calls, transcripts and usage for your own evaluation.

Daily-updated Top 100 Agent Skills rankings — installs, growth, and social buzz aggregated from skills.sh, ClawHub, Tencent SkillHub, GitHub, X and 10+ communities. Open data (CSV).

Guard LLM tool calls with human confirmation and audit logging

A cross-platform OpenAI-compatible proxy for VS Code Chat providers