GH·RADAR

a small, embeddable agent framework in Rust.

难题解答步骤级审计与错误定位器 — 确定性数学引擎，零模型依赖，可离线复跑

Mac menu bar meeting recorder: on-device transcription, live to-dos during the call, notes with Claude or a local model, and your meetings in Claude Code over MCP.

A watchtower for AI agents — scan MCP servers and AI-agent tool manifests for poisoned descriptions, data exfiltration, secrets and supply-chain risk. Zero-dependency, CLI + CI.

Production-shaped MCP server starter: bearer auth, per-client rate limits, host allow-list, structured logging, tests, Docker, CI

Codex 技能：跨对话交接日志。每轮结束自动落盘，让同一项目的多个对话接得上；读取成本恒定，不随项目时长膨胀。

Inspect, manage and update the reference images of iOS snapshot tests (swift-snapshot-testing): a macOS app and CLI.

Classify inbound email, draft replies in your voice, queue for human approval, log everything. Pluggable LLM, offline evals.

Replay terminal output or asciinema casts through a virtual terminal; render text, ANSI or JSON, check, diff

通过 MCP 让 AI 游玩《植物大战僵尸》：读取 JSON 游戏状态，控制暂停与倍速，种植植物、管理存档并显示观战提示。支持 Windows 特定 1.2.0.1073 构建，不包含游戏文件。

Generate conventional commit messages from your staged diff. Free, offline, zero deps. AI mode for $5/mo.

Free, private website analytics for ordinary cPanel hosting. Pageviews, visitors, file downloads and project trends. PHP + SQLite, no cookies, no IPs stored.

Session cost and billing metrics for the DSH Web GUI: what this conversation cost, what the cache saved, and whether the current hour is peak or off-peak. A companion to the shipped stats pills, never a duplicate of them.

A Skills page for the DSH Web GUI: every skill DSH can see, grouped All · Global · Workspaces, with search, copy-command, per-skill origin, and the roots DSH discovers skills from.

Flight recorder for AI agents — Go tracing, deterministic replay from cassettes, and a YAML eval harness with regression gates

Pariksha is a free, open-source exam platform that any school, company or community can run on its own account at zero cost. One deployment hosts many tests. Each test has its own name, ID, URL, question sets, registration form, access rules and proctoring policy. Candidates take tests in a browser; organisers get scores and reports as CSV.

Finds dangerous call sites in LLM / Agent code, and refuses to guess when it cannot tell.

Open-source, self-hosted bookkeeping. Rust + Tauri + React + SQLite, with Plaid/SimpleFIN bank feeds.

No description provided.

Qloo-powered casting agent for disclosed AI characters: brand brief in, ranked characters out, every claim cited to Qloo evidence. Web app + remote MCP server. Qloo Agentic Hackathon.

One entry point for every secrets manager. Run any command with secrets from Infisical, Doppler, Vault, AWS, GCP, Azure, 1Password or Bitwarden.

A reusable agent skill that brings principal-engineer judgment to software design, implementation, code review, debugging, refactoring, migrations, and production readiness—with explicit attention to correctness, security, reliability, architectural trade-offs, and operational cost.

Small tools for running multiple DSH (DeepSeek Harness) sessions at once: a cross-session inbox, a session peek, and a memory-space diagnostic. Multi-session coordination, zero dependencies. Unofficial / 非官方。

Claude Code and Codex debate your architecture, with you as the chair