Aevra CLI — the universal, open-source AI coding agent for your terminal. Local-first, multi-provider, autonomous.
Directory
Browse agent skills
Search by intent: design systems, testing, deployment, security, documentation, Azure, Supabase, React, or any workflow you want an agent to perform better.
Showing 31,111–31,140 of 128,840 skills
Page 1,038 of 4,295 · loading 30 at a time
Deterministic confidence scorecards for RAG systems with signal policy, calibration, evaluator bridges, and index integrity
Render any LaTeX math into crisp, printable PNG/SVG and copy it into Word/WPS/Google Docs in one click — live preview, offline, no LaTeX install. A Claude Code skill that also runs as a standalone web page.
Land all your agents' branches in one command. Pytxo merges your Claude Code, Codex and Copilot branches in a throwaway worktree, runs your tests on the combined result, and lands them only if they pass.
✍️ Turn your sources into polished academic writing — reviews, proposals, essays, SoPs, résumés — every citation cross-checked, not invented. Runs on Claude & Codex. A lite, open-source edition of AcademiCats. 用你的文献写出有据成稿、引用逐条核对 —— 可在 Claude / Codex 上使用。
A MapLibre GL JS plugin for visualizing vector data (GeoJSON, GeoPackage, Shapefile, GeoParquet, FlatGeobuf, CSV) powered by DuckDB-WASM with dynamic vector tiles
🧪 Peer-review your draft across 5 lenses, methodology checked against the right standard per paper type (STROBE/PRISMA/CONSORT/COREQ/Toulmin). Referee letter + scored verdict. Runs on Claude & Codex. A lite, open-source edition of AcademiCats. 按论文类型套用学术标准的模拟同行评审 —— 可在 Claude / Codex 上使用。
🔍 Find real academic papers across 5 scholarly databases & deep-read any open-access PDF — grounded, never fabricated. Runs on Claude & Codex. A lite, open-source edition of AcademiCats. 跨 5 大学术库检索真实论文、精读 PDF。可在 Claude / Codex 上使用。
SeeyonJF-AI chat/completions gateway configuration helper for opencode and Codex.
Two composable Claude Code skills for design-to-code fidelity: extract exact specs from Figma, then verify the running web or mobile app matches them.
Local-first scan for leaked secrets, unsafe defaults, prompt injection, and dangerous agent tools.
Local-first repeatable AI output checks for required content, banned output, schemas, and regressions.
Thin guarded AI API client with retries, Retry-After support, spend caps, request IDs, and receipts.
A React rich-text input that renders {{variable}} tokens as interactive pill chips — built for AI prompt editors, agent builders, and workflow UIs.
Production workflow skills for planning, executing, reviewing, and finishing coding-agent work
Framework-agnostic, realistic highlighter-pen marks for web text: multiline, deterministic, SSR-safe
CLI preflight checks for model VRAM, CUDA, GPU compatibility, and quantization risk.
Unofficial SDK for Codex agents controlling visible ChatGPT web sessions.
Refract any website into its real design system — computed CSS, screenshots, narrative DESIGN.md — in 90 seconds. The verifiable alternative to manually-written design docs.
SAP AI Core Local LLM Proxy - A command-line tool to run a local proxy for SAP AI Core Foundation Models
Claude Code usage monitor — local-first analytics dashboard for Claude Code & Codex CLI sessions. See every session, every project, every dollar. No login, no cloud.
Data client for Reactor JS SDK - PostgREST-style query builder
AI client for Reactor JS SDK - OpenAI-compatible chat completions and embeddings
Strict Goal Mode agents, commands, and guard plugin for OpenCode.
Unofficial PTY-based wrapper for agy (Antigravity CLI / Gemini CLI) - captures Gemini responses that agy writes to the terminal buffer instead of stdout
Switch Pi's AI provider and model based on time of day via YAML time-slot configuration.
Session-scoped /goal command for OpenCode with autonomous execution loop, configurable judge model, and state persistence
Browser-side DOM + WebMCP event observer for AgentForge. Captures user-intent signals (clicks, navigation, inputs hashed, scroll) and exposes them through a transport-agnostic drain() buffer. Designed to plug into the @agentforge-io/chat-sdk session or th
AI-powered i18n automation CLI for React — scan, translate, and transform hardcoded strings into react-i18next calls
AIACA — AI Agentic Coding Assistant. A production-grade agentic CLI with dual-provider support (Gemini + Claude), rich TUI, and 100+ tools.