diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml index 0822fb6..ada282a 100644 --- a/.github/workflows/ci.yml +++ b/.github/workflows/ci.yml @@ -59,5 +59,14 @@ jobs: - name: Lychee link checker uses: lycheeverse/lychee-action@v2 with: - args: --no-progress --exclude-loopback --exclude 'github.com/lglucas/boringco' '**/*.md' + args: >- + --no-progress + --exclude-loopback + --exclude 'github\.com/lglucas/boringco' + --exclude 'linkedin\.com' + --exclude 'lume\.ufrgs\.br' + --exclude 'perestroika\.com\.br' + --max-retries 3 + --retry-wait-time 5 + '**/*.md' fail: true diff --git a/docs/registry/INDEX.md b/docs/registry/INDEX.md index c498bbb..faaa9f9 100644 --- a/docs/registry/INDEX.md +++ b/docs/registry/INDEX.md @@ -6,7 +6,7 @@ For how the registry works, see [`README.md`](README.md). For per-pack detail, click the slug in the first column. For non-installable websites and live tools, see the **External resources** section below. -Last index update: 2026-05-01 (v0.4.3 — quick wins: ETHOS + 3 skills + agents-marketplace tag) +Last index update: 2026-05-09 (registry-additions-2026-05-09 — 14 new packs added, 3 refreshed) --- @@ -14,31 +14,41 @@ Last index update: 2026-05-01 (v0.4.3 — quick wins: ETHOS + 3 skills + agents- | Slug | Name | Tags | Status | License | Last reviewed | |------|------|------|--------|---------|---------------| +| [`agent-exchange`](packs/agent-exchange.md) | Agent Exchange (AEX) | `agents` `infra` `experimental` | experimental | MIT | 2026-05-09 | +| [`agent-skills-eval`](packs/agent-skills-eval.md) | agent-skills-eval | `agents` `tooling` | active | MIT | 2026-05-09 | | [`agentic-inbox`](packs/agentic-inbox.md) | Agentic Inbox (Cloudflare) | `productivity` `agents` `stack-pack` | active | Apache 2.0 | 2026-04-30 | | [`anthropics-skills`](packs/anthropics-skills.md) | Anthropic Skills | `foundations` | active | MIT | 2026-04-30 | | [`antigravity-awesome-skills`](packs/antigravity-awesome-skills.md) | Antigravity Awesome Skills | `ecosystem` `agents` `tooling` | active | MIT + CC BY 4.0 | 2026-04-30 | | [`app-store-screenshots`](packs/app-store-screenshots.md) | App Store Screenshots | `design` `tooling` | active | MIT | 2026-04-30 | -| [`auto-research-claw`](packs/auto-research-claw.md) | AutoResearchClaw | `ai` `research` `agents` | active | MIT | 2026-04-30 | +| [`ascii-draw`](packs/ascii-draw.md) | ASCII Draw | `design` `tooling` | active | GPL-3.0 | 2026-05-09 | +| [`auto-research-claw`](packs/auto-research-claw.md) | AutoResearchClaw | `ai` `research` `agents` | active | MIT | 2026-05-09 | | [`autohedge`](packs/autohedge.md) | AutoHedge | `experimental` `fintech` `agents` | experimental | MIT | 2026-04-30 | | [`awesome-design-md`](packs/awesome-design-md.md) | VoltAgent awesome-design-md | `design` | active | check upstream | 2026-04-30 | | [`bark`](packs/bark.md) | Bark (Suno) | `ai` `audio` | active | MIT | 2026-04-30 | | [`boringco`](packs/boringco.md) | Boring Co (origin) | `origin` | archived | private | 2026-04-30 | | [`cis-hardening-guide`](packs/cis-hardening-guide.md) | CIS Hardening Guide | `security` `infra` | active | check upstream | 2026-04-30 | +| [`claude-code-harness`](packs/claude-code-harness.md) | Claude Code Harness | `ecosystem` `tooling` | active | MIT | 2026-05-09 | | [`claude-cookbooks`](packs/claude-cookbooks.md) | Anthropic Claude Cookbooks | `cookbook` `foundations` `ai` | active | MIT | 2026-04-30 | | [`clip`](packs/clip.md) | CLIP (OpenAI) | `ai` `vision` | active | MIT | 2026-04-30 | +| [`crawlee`](packs/crawlee.md) | Crawlee (Apify) | `tooling` `ai` `data-engineering` | active | Apache-2.0 | 2026-05-09 | +| [`cryptpad`](packs/cryptpad.md) | CryptPad | `productivity` `security` | active | AGPL-3.0 | 2026-05-09 | | [`data-engineer-handbook`](packs/data-engineer-handbook.md) | Data Engineer Handbook | `learning` `data-engineering` | active | check upstream | 2026-04-30 | | [`data-engineering-zoomcamp`](packs/data-engineering-zoomcamp.md) | DE Zoomcamp (DataTalksClub) | `learning` `data-engineering` | active | check upstream | 2026-04-30 | | [`diagram-design`](packs/diagram-design.md) | Diagram Design (Cathryn Lavery) | `design` `ai` `tooling` | active | check upstream | 2026-04-30 | | [`everything-claude-code`](packs/everything-claude-code.md) | everything-claude-code | `ecosystem` `tooling` | active | check upstream | 2026-04-30 | -| [`fincept-terminal`](packs/fincept-terminal.md) | Fincept Terminal | `fintech` `tooling` | active | AGPL-3.0 + commercial | 2026-04-30 | +| [`fincept-terminal`](packs/fincept-terminal.md) | Fincept Terminal | `fintech` `tooling` | active | AGPL-3.0 + commercial | 2026-05-09 | | [`fireworks-tech-graph`](packs/fireworks-tech-graph.md) | Fireworks Tech Graph | `design` `ai` `tooling` | active | MIT | 2026-04-30 | | [`gitleaks`](packs/gitleaks.md) | Gitleaks | `security` `tooling` | active | MIT | 2026-04-30 | | [`google-research`](packs/google-research.md) | Google Research | `ai` `ml` `research` | active | Apache 2.0 + CC BY 4.0 | 2026-04-30 | +| [`googleworkspace-cli`](packs/googleworkspace-cli.md) | Google Workspace CLI (`gws`) ⭐ | `agents` `productivity` `tooling` `agents-marketplace` | active | Apache-2.0 | 2026-05-09 | | [`gstack`](packs/gstack.md) | gstack ⭐ | `ecosystem` `agents` `foundations` | active | MIT | 2026-04-30 | | [`holyclaude`](packs/holyclaude.md) | HolyClaude | `ecosystem` `tooling` `stack-pack` | active | MIT | 2026-04-30 | | [`kronos`](packs/kronos.md) | Kronos (Financial Foundation Model) | `ai` `ml` `fintech` `research` | active | MIT | 2026-04-30 | | [`hyperframes`](packs/hyperframes.md) | HyperFrames | `ai` `design` `tooling` | active | Apache 2.0 | 2026-04-30 | +| [`isomiddleearth`](packs/isomiddleearth.md) | Iso Middle Earth | `design` `tooling` | active | check upstream | 2026-05-09 | +| [`jeweledtech-agentic-framework`](packs/jeweledtech-agentic-framework.md) | JeweledTech Agentic Framework | `agents` `experimental` | active | MIT | 2026-05-09 | | [`llms-from-scratch`](packs/llms-from-scratch.md) | LLMs from Scratch (Raschka) | `learning` `ai` | active | see upstream | 2026-04-30 | +| [`locally-uncensored`](packs/locally-uncensored.md) | Locally Uncensored | `ai` `tooling` | active | AGPL-3.0 | 2026-05-09 | | [`made-with-ml`](packs/made-with-ml.md) | Made With ML | `learning` `ml` | active | MIT | 2026-04-30 | | [`microsoft-ai-agents-for-beginners`](packs/microsoft-ai-agents-for-beginners.md) | MS AI Agents for Beginners | `learning` `agents` `ai` | active | MIT | 2026-04-30 | | [`microsoft-ai-for-beginners`](packs/microsoft-ai-for-beginners.md) | MS AI for Beginners | `learning` `ai` | active | MIT | 2026-04-30 | @@ -46,16 +56,19 @@ Last index update: 2026-05-01 (v0.4.3 — quick wins: ETHOS + 3 skills + agents- | [`microsoft-generative-ai-for-beginners`](packs/microsoft-generative-ai-for-beginners.md) | MS Generative AI for Beginners | `learning` `ai` | active | MIT | 2026-04-30 | | [`microsoft-ml-for-beginners`](packs/microsoft-ml-for-beginners.md) | MS ML for Beginners | `learning` `ml` | active | MIT | 2026-04-30 | | [`miroshark`](packs/miroshark.md) | MiroShark | `ai` `agents` `research` `experimental` | active | AGPL-3.0 | 2026-04-30 | +| [`mirothinker`](packs/mirothinker.md) | MiroThinker | `ai` `agents` `research` | active | Apache-2.0 | 2026-05-09 | | [`nellavio`](packs/nellavio.md) | Nellavio ⭐ | `stack-pack` `design` | active | MIT | 2026-04-30 | | [`open-design`](packs/open-design.md) | Open Design (Nexu) | `design` `ai` `ecosystem` | active | Apache-2.0 | 2026-04-30 | | [`open-higgsfield-ai`](packs/open-higgsfield-ai.md) | Open Higgsfield AI | `ai` `vision` `experimental` | active | MIT | 2026-04-30 | | [`openai-cookbook`](packs/openai-cookbook.md) | OpenAI Cookbook | `cookbook` `foundations` `ai` | active | MIT | 2026-04-30 | +| [`openhuman`](packs/openhuman.md) | OpenHuman | `agents` `productivity` `tooling` | active beta | GPL-3.0 | 2026-05-09 | | [`optio`](packs/optio.md) | Optio (Jonathan Wiggins) | `ecosystem` `agents` `tooling` `infra` | active | MIT | 2026-04-30 | | [`pathway-llm-app`](packs/pathway-llm-app.md) | Pathway LLM App | `ai` `data-engineering` `tooling` | active | MIT | 2026-04-30 | | [`phantom-ui`](packs/phantom-ui.md) | Phantom-UI (skeleton loaders) | `design` `tooling` | active | MIT | 2026-04-30 | | [`python-100-days`](packs/python-100-days.md) | Python-100-Days | `learning` | active | check upstream | 2026-04-30 | | [`python-data-science-handbook`](packs/python-data-science-handbook.md) | Python DS Handbook (VanderPlas) | `learning` `ml` | active | MIT + CC-BY-NC-ND | 2026-04-30 | -| [`ruflo`](packs/ruflo.md) | Ruflo | `ecosystem` `agents` `tooling` | active | MIT | 2026-04-30 | +| [`quran-database`](packs/quran-database.md) | Quran Database | `learning` `research` `dataset` | active | check upstream | 2026-05-09 | +| [`ruflo`](packs/ruflo.md) | Ruflo | `ecosystem` `agents` `tooling` | active | MIT | 2026-05-09 | | [`segment-anything`](packs/segment-anything.md) | Segment Anything (SAM) | `ai` `vision` | active | Apache 2.0 | 2026-04-30 | | [`slavingia-skills`](packs/slavingia-skills.md) | Slavingia Skills ⭐ (Minimalist Entrepreneur) | `foundations` `learning` `ecosystem` | active | check upstream | 2026-04-30 | | [`solana-claude`](packs/solana-claude.md) | Solana Claude | `stack-pack` `web3` | active | check upstream | 2026-04-30 | @@ -67,6 +80,7 @@ Last index update: 2026-05-01 (v0.4.3 — quick wins: ETHOS + 3 skills + agents- | [`tradingagents`](packs/tradingagents.md) | TradingAgents (Tauric) ⭐ | `fintech` `agents` `ai` `research` | active | Apache-2.0 | 2026-04-30 | | [`trailofbits-skills`](packs/trailofbits-skills.md) | Trail of Bits Skills | `security` | active | check upstream | 2026-04-30 | | [`ui-ux-pro-max`](packs/ui-ux-pro-max.md) | UI UX Pro Max Skill | `design` | active | check upstream | 2026-04-30 | +| [`understand-anything`](packs/understand-anything.md) | Understand-Anything | `tooling` `productivity` `ai` `ecosystem` | active | MIT | 2026-05-09 | | [`unity-mcp`](packs/unity-mcp.md) | Unity-MCP | `stack-pack` `agents` `tooling` | active | Apache-2.0 | 2026-04-30 | | [`vibeyard`](packs/vibeyard.md) | Vibeyard | `ecosystem` `productivity` `tooling` | active | MIT | 2026-04-30 | | [`vibe-trading`](packs/vibe-trading.md) | Vibe-Trading | `fintech` `agents` `ai` | active | MIT | 2026-04-30 | diff --git a/docs/registry/packs/agent-exchange.md b/docs/registry/packs/agent-exchange.md new file mode 100644 index 0000000..0ba4913 --- /dev/null +++ b/docs/registry/packs/agent-exchange.md @@ -0,0 +1,55 @@ +# Agent Exchange (AEX) + +- **URL:** https://github.com/open-experiments/agent-exchange +- **License:** MIT +- **Status:** experimental (402 stars) +- **Last reviewed:** 2026-05-09 +- **Tags:** #agents #infra #experimental + +## What it is + +A programmatic **marketplace for AI agents** that applies ad-tech economics (auction, bidding, settlement) to inter-agent work. Solves the "N×M integration problem" — instead of every consumer agent integrating with every provider agent point-to-point, AEX is a broker where agents discover each other, bid on work, and settle payments automatically. + +Implemented as 10 Go microservices on MongoDB + Docker + GCP Cloud Run, with a Python + NiceGUI demo UI. + +## When to install + +- Designing a multi-vendor agent ecosystem where buying-side and selling-side scale independently. +- Researching agent economics, bid mechanisms, and trust frameworks. +- Building an internal agent marketplace where teams expose capabilities and other teams consume them with metering. +- Need a reference implementation of agent-to-agent settlement. + +## When NOT to install + +- Single-agent, single-team use cases — wrong tool, this is infra-heavy. +- Need a production-ready marketplace today — explicitly experimental, treat as a research substrate. +- Stack doesn't allow Go services + MongoDB at the infra layer. + +## How to install + +```bash +git clone https://github.com/open-experiments/agent-exchange +cd agent-exchange +docker compose up +# 10 services, give it time +``` + +## Fit signals + +- Project explores agent-to-agent commerce, billing, or settlement. +- Multiple teams or vendors will publish/consume agent capabilities. +- Studying ad-tech-style auction mechanisms applied to AI. +- Interested in the `everything-claude-code:agent-payment-x402` pattern at marketplace scale. + +## Conflicts and overlaps + +- Conceptual overlap with `agents-marketplace` tag — but those packs are **bundled skill collections** (single-author marketplaces); AEX is a **broker for agent transactions** (multi-vendor exchange). +- Pairs with `everything-claude-code:agent-payment-x402` (per-agent budgets and wallets) and `everything-claude-code:enterprise-agent-ops` (long-lived workloads). +- Composes with `mirothinker` or any agent that wants to sell or buy work through an exchange. + +## Notes + +- MIT — safe for commercial use. +- ⚠️ **Experimental** — APIs, schemas, and economic mechanisms may change. Do not build production billing on top of it without a fork. +- 10 microservices is real infra weight; not a "vibe-coder MVP" choice. +- Worth studying as a reference even if you don't deploy — the economic model and protocol design are well thought out. diff --git a/docs/registry/packs/agent-skills-eval.md b/docs/registry/packs/agent-skills-eval.md new file mode 100644 index 0000000..d538d08 --- /dev/null +++ b/docs/registry/packs/agent-skills-eval.md @@ -0,0 +1,54 @@ +# agent-skills-eval + +- **URL:** https://github.com/darkrishabh/agent-skills-eval +- **License:** MIT +- **Status:** active (257 stars) +- **Last reviewed:** 2026-05-09 +- **Tags:** #agents #tooling + +## What it is + +A test runner for `agentskills.io`-style AI agent skills. Empirically validates whether a given skill actually improves model performance: runs the same prompt **with** and **without** the skill loaded, then uses a judge model to grade outputs and produce evidence-backed reports on skill effectiveness. + +Built for the Agent Skills ecosystem (Claude Skills, Anthropic Skills format) but works with any OpenAI-compatible API. + +## When to install + +- Authoring custom skills (Claude Code, Anthropic Skills, agentskills.io) and need objective evidence they help. +- Comparing two skill versions before promoting one as the canonical version. +- CI gate for a skill repo — fail PRs that regress skill quality. +- Pre-release validation before publishing a skill bundle to a marketplace. + +## When NOT to install + +- No custom skills being authored — this is for skill builders, not consumers. +- Cost-sensitive environments without judge-model budget — judge runs cost API tokens. +- Subjective output domains (creative writing, design) where judge-model grading is unreliable. + +## How to install + +```bash +git clone https://github.com/darkrishabh/agent-skills-eval +cd agent-skills-eval +npm install +# Configure provider (OpenAI-compatible) and judge model +``` + +## Fit signals + +- Project includes `.claude/skills/`, `agents/`, or `skills/` directories you wrote yourself. +- Team publishes skills externally and wants quality evidence. +- Need to A/B test prompt/skill variants empirically. + +## Conflicts and overlaps + +- Complementary to `everything-claude-code:eval-harness` — formal eval framework for sessions; agent-skills-eval is narrower (single-skill A/B). +- Complementary to `everything-claude-code` skill management commands (`skill-create`, `skill-stocktake`, `skill-comply`, `skill-health`) — those manage skills; this validates them. +- Pairs with `multi-ai-review` skill in this OS for hard decisions about which skill version to ship. + +## Notes + +- MIT — safe for commercial use. +- 257 stars — niche but high-leverage tool. The skill-authoring user base is small but growing fast. +- Critical for `agents-marketplace` packs that want to claim quality — without evals, "this skill helps" is just a vibe. +- TypeScript / Node — minimal infra to run. diff --git a/docs/registry/packs/ascii-draw.md b/docs/registry/packs/ascii-draw.md new file mode 100644 index 0000000..f16fbc2 --- /dev/null +++ b/docs/registry/packs/ascii-draw.md @@ -0,0 +1,55 @@ +# ASCII Draw + +- **URL:** https://github.com/Nokse22/ascii-draw +- **License:** GPL-3.0 +- **Status:** active (462 stars) +- **Last reviewed:** 2026-05-09 +- **Tags:** #design #tooling + +## What it is + +A desktop drawing application (Python + GTK4 + libadwaita) for creating diagrams, flowcharts, tables, organizational charts, and ASCII art using only character glyphs. Bundled with `pyfiglet` for figure-text and `emoji` for inline glyphs. Targets Linux desktop primarily but the Python/GTK stack is portable. + +Useful for embedding clean ASCII diagrams in code comments, READMEs, technical docs, and CLI help text — without resorting to image attachments. + +## When to install + +- Authoring technical docs / READMEs that benefit from inline ASCII diagrams. +- Generating flowcharts that need to live in plaintext (terminal apps, plain markdown, email). +- Need an offline GUI alternative to ASCIIFlow or web-based ASCII-diagram tools. +- Linux-first dev environment with GTK4 already installed. + +## When NOT to install + +- Need polished publication-grade diagrams — use Mermaid, Excalidraw, or `diagram-design` instead. +- Windows-only environment without GTK4 — install path is rougher. +- Diagrams need to render in a browser anyway — use a browser-native tool. + +## How to install + +```bash +# Linux (preferred) +flatpak install flathub io.github.nokse22.asciidraw +# or via source +git clone https://github.com/Nokse22/ascii-draw +``` + +## Fit signals + +- Project documentation lives in plain markdown / text files. +- Codebase has CLI tools whose `--help` output benefits from ASCII layout diagrams. +- Team prefers GUI canvas over markdown-text diagram syntax for layout work. +- Already on a GNOME / GTK desktop. + +## Conflicts and overlaps + +- Functional overlap with `tegaki` (drawing) — different output (raster vs ASCII). +- Functional overlap with `diagram-design` (Cathryn Lavery editorial diagrams) — different fidelity (plaintext vs publication-grade). +- Functional overlap with `fireworks-tech-graph` (NL-to-SVG) — different format (ASCII vs SVG). + +## Notes + +- ⚠️ **GPL-3.0** — fine for personal / internal use; review before bundling output-generation code in a commercial product. +- The output (ASCII art itself) is yours — only the application is GPL. +- 462 stars — small but loyal user base. +- A surprisingly useful authoring tool for technical writing if your docs are plaintext-first. diff --git a/docs/registry/packs/auto-research-claw.md b/docs/registry/packs/auto-research-claw.md index 557903f..19dcb9c 100644 --- a/docs/registry/packs/auto-research-claw.md +++ b/docs/registry/packs/auto-research-claw.md @@ -2,8 +2,8 @@ - **URL:** https://github.com/aiming-lab/AutoResearchClaw - **License:** MIT -- **Status:** active (11.8k stars, v0.4.0 April 2026, 235 commits, 2,699 tests passing) -- **Last reviewed:** 2026-04-30 +- **Status:** active (12k stars, v0.4.0 April 2026 — Human-in-the-Loop Co-Pilot) +- **Last reviewed:** 2026-05-09 - **Tags:** #ai #research #agents ## What it is diff --git a/docs/registry/packs/claude-code-harness.md b/docs/registry/packs/claude-code-harness.md new file mode 100644 index 0000000..b02075b --- /dev/null +++ b/docs/registry/packs/claude-code-harness.md @@ -0,0 +1,55 @@ +# Claude Code Harness + +- **URL:** https://github.com/chachamaru127/claude-code-harness +- **License:** MIT +- **Status:** active (806 stars) +- **Last reviewed:** 2026-05-09 +- **Tags:** #ecosystem #tooling + +## What it is + +A dedicated development harness for Claude Code that enforces a structured **Plan → Work → Review → Release** cycle with runtime safety guardrails, parallel task execution, and autonomous code review from multiple perspectives. The v4 "Hokage" release replaced the Node.js dependency with a **Go-native guardrail engine** for performance and packaging. + +Effectively turns Claude Code into a disciplined development partner instead of a free-form coding chat. + +## When to install + +- Want enforced workflow discipline (plan before code, review after code) without remembering to invoke skills manually. +- Need parallel task execution managed by a harness (instead of ad-hoc multiple Claude Code sessions). +- Looking for autonomous multi-perspective code review (security, perf, idiom checks) on every change. +- Strong preference for Go-native binaries over Node.js dependency surface. + +## When NOT to install + +- First-time Claude Code user — start with the basics; harnesses add cognitive load. +- Already using `gstack`, `holyclaude`, or another orchestration pack — choose one harness, not two. +- Highly exploratory / spike work where rigid workflow gates slow you down. + +## How to install + +```bash +git clone https://github.com/chachamaru127/claude-code-harness +cd claude-code-harness +# Follow upstream README — Go binary install replaces previous Node setup +``` + +## Fit signals + +- Team already runs Claude Code daily and wants more workflow discipline. +- Project benefits from enforced plan-before-code (regulated, security-sensitive, high-impact systems). +- Multi-language codebase where guardrails need to be language-agnostic at the engine level. + +## Conflicts and overlaps + +- Overlaps with `gstack` (23-agent virtual team) — gstack is agent-bundle-centric; this harness is workflow-engine-centric. +- Overlaps with `holyclaude` (containerized Claude Code env with 7 AI CLIs) — different layer (env vs workflow). +- Overlaps with `optio` (K8s ticket-to-PR orchestration) — Optio is infra-heavy; this is local-dev-focused. +- Overlaps with `everything-claude-code:autonomous-loops` and `gan-style-harness` skills — those are patterns; this is a packaged harness. +- See [`tags/agents-marketplace.md`](../tags/agents-marketplace.md) "Conflict resolution" guidance — pick one harness per project. + +## Notes + +- MIT — safe for commercial use and modification. +- 806 stars and growing — Go-native rewrite is a meaningful quality signal. +- Author handle: `chachamaru127` — verify identity if vendor-trust matters. +- v4 "Hokage" is the recommended baseline as of last review. diff --git a/docs/registry/packs/crawlee.md b/docs/registry/packs/crawlee.md new file mode 100644 index 0000000..9b76eeb --- /dev/null +++ b/docs/registry/packs/crawlee.md @@ -0,0 +1,54 @@ +# Crawlee + +- **URL:** https://github.com/apify/crawlee +- **License:** Apache-2.0 +- **Status:** active (23.1k stars) +- **Last reviewed:** 2026-05-09 +- **Tags:** #tooling #ai #data-engineering + +## What it is + +Production-grade web scraping and browser automation library for Node.js. Builds reliable crawlers in JavaScript / TypeScript with first-class support for Puppeteer, Playwright, Cheerio, and JSDOM. Designed explicitly for AI / LLM / RAG / GPT data pipelines — extracts HTML, PDF, images, and other files while evading common bot-detection patterns. + +Maintained by Apify (the company behind the eponymous scraping platform), so the open-source library tracks production scraping needs at scale. + +## When to install + +- Building a RAG pipeline that needs fresh web data (news, docs, prices, reviews). +- LLM / agent ingestion pipelines that scrape, normalize, and feed into vector stores. +- Need anti-detection scraping with proxy rotation, fingerprinting, and queue management out of the box. +- Replacing brittle in-house scrapers with a battle-tested library. + +## When NOT to install + +- One-off scraping of a single page — `fetch()` + `cheerio` directly is enough. +- Only need API integrations (no scraping) — wrong tool. +- Compliance constraints prohibit scraping the target site (TOS, copyright, robots.txt) — Crawlee makes scraping easier, not legal. + +## How to install + +```bash +npm install crawlee playwright +# or +npx crawlee create my-crawler +``` + +## Fit signals + +- Project includes a knowledge ingestion or research-agent component. +- Stack uses Node.js / TypeScript. +- Need scheduled scraping with retries, queues, and dataset persistence. +- `pathway-llm-app` or any RAG pack already in scope and needs a scraping front-end. + +## Conflicts and overlaps + +- Composes with `pathway-llm-app` (RAG) — Crawlee feeds, Pathway indexes. +- Composes with `mirothinker` and `auto-research-claw` (research agents) as a tool source. +- Overlaps with the `data-scraper-agent` skill in `everything-claude-code` — that skill builds a scheduled scraping agent; Crawlee is the lower-level engine. + +## Notes + +- Apache-2.0 — safe for commercial use with attribution. +- 23.1k stars — one of the most established scraping libraries in the Node ecosystem. +- ⚠️ Scraping legality varies by jurisdiction and target — always check robots.txt, TOS, and copyright before deploying production crawlers. +- Apify Cloud is the hosted commercial offering — the OSS library does not require it. diff --git a/docs/registry/packs/cryptpad.md b/docs/registry/packs/cryptpad.md new file mode 100644 index 0000000..3507b55 --- /dev/null +++ b/docs/registry/packs/cryptpad.md @@ -0,0 +1,58 @@ +# CryptPad + +- **URL:** https://github.com/cryptpad/cryptpad +- **License:** AGPL-3.0 +- **Status:** active (7.6k stars) +- **Last reviewed:** 2026-05-09 +- **Tags:** #productivity #security + +## What it is + +A complete **end-to-end encrypted collaborative office suite**: real-time collaborative documents, spreadsheets, presentations, forms, whiteboards, and code editors. The server only sees encrypted blobs — even a compromised server cannot read user content. Self-hostable with Docker. + +Effectively the privacy-first answer to Google Workspace / Office 365 — collaboration without trusting the host with plaintext. + +## When to install + +- Privacy-sensitive teams (legal, medical, journalism, activism, regulated industries). +- Self-hosted alternative to Google Docs / Office that the team can audit. +- Collaboration where the threat model includes "the server operator" (insider risk). +- Studying a working real-world implementation of E2E-encrypted real-time collaboration. + +## When NOT to install + +- Need every Google/Office feature on day one — feature parity is not the goal. +- Mobile-first workflows — desktop browser is the primary surface. +- Cannot self-host and don't want to use the public CryptPad instance. +- AGPL-3.0 conflicts with your distribution model. + +## How to install + +```bash +git clone https://github.com/cryptpad/cryptpad +cd cryptpad +docker compose up +# or follow upstream for native install +``` + +Or use the public hosted instance at [https://cryptpad.fr](https://cryptpad.fr) for personal use. + +## Fit signals + +- Project handles confidential client data and needs E2E-encrypted collaboration. +- Compliance regime requires demonstrable data minimization at the host level. +- Threat model includes server-operator compromise. +- Team is comfortable self-hosting Node.js services. + +## Conflicts and overlaps + +- Functional overlap with Google Workspace / Office 365 — but trust model is fundamentally different (E2E encryption). +- Composes with `cis-hardening-guide` (server hardening for the self-host). +- See [`tags/security.md`](../tags/security.md) for related packs. + +## Notes + +- ⚠️ **AGPL-3.0** — strong copyleft, network-use clause applies if you self-host with modifications. +- E2E encryption is real, not marketing — verified by external audits in the past. Still: review threat model assumptions for your specific use case. +- Active maintenance, well-known project (XWiki SAS / French security ecosystem origins). +- Performance scales well for small teams; large-org deployments need careful sizing. diff --git a/docs/registry/packs/fincept-terminal.md b/docs/registry/packs/fincept-terminal.md index c2b90f6..28a7bdd 100644 --- a/docs/registry/packs/fincept-terminal.md +++ b/docs/registry/packs/fincept-terminal.md @@ -2,8 +2,8 @@ - **URL:** https://github.com/Fincept-Corporation/FinceptTerminal - **License:** AGPL-3.0 (open source / personal) + commercial license required for business/SaaS/forks -- **Status:** active (18.3k stars, v4.0.2 April 2026, 26 releases) -- **Last reviewed:** 2026-04-30 +- **Status:** active (20.5k stars, v4.0.2 April 2026, 26 releases) +- **Last reviewed:** 2026-05-09 - **Tags:** #fintech #tooling ## What it is diff --git a/docs/registry/packs/googleworkspace-cli.md b/docs/registry/packs/googleworkspace-cli.md new file mode 100644 index 0000000..f0aca52 --- /dev/null +++ b/docs/registry/packs/googleworkspace-cli.md @@ -0,0 +1,55 @@ +# Google Workspace CLI (`gws`) + +- **URL:** https://github.com/googleworkspace/cli +- **License:** Apache-2.0 +- **Status:** active (26k stars) +- **Last reviewed:** 2026-05-09 +- **Tags:** #agents #productivity #tooling #agents-marketplace + +## What it is + +The **official Google Workspace CLI** — a unified command-line tool covering Drive, Gmail, Calendar, Sheets, Docs, Chat, and Admin APIs. Built explicitly for "humans and AI agents" with structured JSON output, auto-pagination, and **100+ bundled `SKILL.md` files** (40+ for Gmail/Drive/Docs/Calendar/Sheets workflows alone). + +Written in Rust (98%) for fast startup and clean install. Distributed via npm and Nix flakes. + +## When to install + +- Project automates Google Workspace and you don't want to wire 6 separate Google APIs by hand. +- AI agent workflows that read/write Workspace content (email triage, doc generation, calendar ops). +- Want a curated `SKILL.md` library you can drop into Claude Code / Cursor / Copilot CLI. +- Need stable, structured JSON output instead of brittle screen-scraping. + +## When NOT to install + +- Project has no Google Workspace surface. +- Need server-side service-account flows only (CLI is human-auth-first; check OAuth model fits). +- Strict enterprise IT policy prohibits third-party CLIs over Workspace data. + +## How to install + +```bash +npm install -g @googleworkspace/cli +gws auth login +``` + +Or via Nix flake — see upstream README. + +## Fit signals + +- Project includes `everything-claude-code:google-workspace-ops` skill or similar Workspace automation. +- Replacing brittle Apps Script / Zapier / point-tool integrations with a stable CLI. +- Building an inbox-triage, doc-pipeline, or calendar-bot agent. + +## Conflicts and overlaps + +- Composes with `everything-claude-code:google-workspace-ops`, `chief-of-staff` agent, `email-ops` skill — `gws` is the underlying engine, those skills are the orchestration layer. +- Overlaps with `agentic-inbox` (Cloudflare-based email triage) — different host (Cloudflare Workers vs local CLI). +- Strong `agents-marketplace` candidate because of the bundled 100+ SKILL.md files — see [`tags/agents-marketplace.md`](../tags/agents-marketplace.md). + +## Notes + +- ✅ **Official Google Workspace tool** — strong vendor signal, long-term maintenance expected. +- Apache-2.0 — safe for commercial use with attribution. +- 26k stars — top-tier ecosystem signal. +- The "built for AI agents" framing is real: structured output, deterministic flags, no interactive prompts. +- Use as the primary integration layer instead of writing custom Google API client code. diff --git a/docs/registry/packs/isomiddleearth.md b/docs/registry/packs/isomiddleearth.md new file mode 100644 index 0000000..b041f0e --- /dev/null +++ b/docs/registry/packs/isomiddleearth.md @@ -0,0 +1,51 @@ +# Iso Middle Earth + +- **URL:** https://github.com/hasanharman/isomiddleearth +- **License:** check upstream (not explicitly stated) +- **Status:** reference-only until upstream license is explicit (553 stars) +- **Last reviewed:** 2026-05-09 +- **Tags:** #design #tooling + +## What it is + +An interactive isometric world-builder set in Tolkien's Middle-earth. Users design custom landscapes across seven realms by placing terrain tiles and character overlays, with export. Built as a Next.js 16 + React 19 + TypeScript + Tailwind app using Radix UI primitives and Zustand state management. + +Useful as a **reference implementation** of a polished interactive Next.js app with a tile-based editor and Radix UI patterns — beyond its pure entertainment value. + +## When to install + +- Studying a clean Next.js 16 / React 19 / Zustand reference implementation. +- Building a tile-based editor or any grid-canvas UI — copy patterns rather than reinvent. +- Need inspiration for a polished isometric or game-like web UI. +- Looking for Radix UI + Tailwind composition patterns at scale. + +## When NOT to install + +- No Next.js / React in the stack. +- Looking for a generic art tool — this is Tolkien-themed; assets and lore are domain-specific. +- License clarity matters and you cannot wait for upstream clarification. + +## How to install + +```bash +git clone https://github.com/hasanharman/isomiddleearth +cd isomiddleearth +pnpm install && pnpm dev +``` + +## Fit signals + +- `package.json` includes `next@16`, `react@19`, `zustand`, `@radix-ui/*`, `tailwindcss@4`. +- Project needs an interactive grid/tile editor. +- Designer / dev wants a strong reference for "polished isometric web UI." + +## Conflicts and overlaps + +- Pure visual/UX inspiration overlap with `tegaki` (drawing) and `ascii-draw` (ASCII canvas). +- Reference-implementation overlap with `nellavio` (stack-pack) — Nellavio gives a stack template, IsoMiddleEarth shows a deployed app at scale. + +## Notes + +- ⚠️ **License not explicitly stated** in README at last review. Treat as "view-only reference" until upstream clarifies — do not copy-paste production code without confirming a permissive license. +- Tolkien IP — Middle-earth assets and naming are Estate-protected; do not lift visuals into commercial work. +- Strong tech-stack fit if you're already on the Next.js 16 / React 19 / Tailwind 4 / Zustand wave. diff --git a/docs/registry/packs/jeweledtech-agentic-framework.md b/docs/registry/packs/jeweledtech-agentic-framework.md new file mode 100644 index 0000000..eb64392 --- /dev/null +++ b/docs/registry/packs/jeweledtech-agentic-framework.md @@ -0,0 +1,55 @@ +# JeweledTech Agentic Framework + +- **URL:** https://github.com/jeweledtech/agentic-framework +- **License:** MIT +- **Status:** active (17 stars) +- **Last reviewed:** 2026-05-09 +- **Tags:** #agents #experimental + +## What it is + +An orchestration-agnostic framework for scaffolding multi-agent business systems — "operate a fully functional company through a hierarchy of AI agents covering sales, marketing, engineering, customer service, back-office, and security." Ships with 15+ pre-built agents, 87 n8n workflows, FastAPI endpoints, and a demo UI. Supports Anthropic Claude, OpenAI, Ollama, and LiteLLM as LLM providers. + +Targets entrepreneurs and SMBs that want to operate a "digital workforce" without integrating dozens of point tools. + +## When to install + +- Solo founder or small team wants a starting scaffold for multi-department AI ops. +- Already running n8n and want a curated 87-workflow head start instead of building from zero. +- Researching how to organize agents into business departments (sales, eng, support). +- Need a demoable "AI company in a box" for client pitches or workshops. + +## When NOT to install + +- Single-agent or single-department use case — overkill. +- Production-grade multi-tenant SaaS — this is solo / SMB scope. +- Strict compliance environment without time to audit each of the 87 workflows. +- 17 stars suggests early-stage adoption — verify maintenance cadence before betting on it. + +## How to install + +```bash +git clone https://github.com/jeweledtech/agentic-framework +cd agentic-framework +# Follow upstream — FastAPI + n8n + Ollama or cloud provider +docker compose up +``` + +## Fit signals + +- Project is "build a company / agency / SMB ops layer" rather than "build a single feature." +- Team uses n8n already. +- Comfort with Python (89% of codebase) and FastAPI. + +## Conflicts and overlaps + +- Conceptual overlap with `gstack` (23-agent virtual engineering team) — gstack is **engineering-only**; this framework is **business-wide**. +- Conceptual overlap with `optio` (K8s ticket-to-PR workflow) — Optio is engineering ops; this covers wider business ops. +- Composes with `agent-exchange` (AEX) if multiple of these "digital workforces" need to transact. + +## Notes + +- MIT — safe for commercial use. +- ⚠️ **Low star count (17)** at last review — early-stage adoption signal. Treat as experimental and verify activity before relying on it. +- 87 n8n workflows is significant scaffolding — even partial reuse may justify the install. +- Strong fit for `vibe-coders` building business automation; weaker fit for production engineering teams. diff --git a/docs/registry/packs/locally-uncensored.md b/docs/registry/packs/locally-uncensored.md new file mode 100644 index 0000000..703d0be --- /dev/null +++ b/docs/registry/packs/locally-uncensored.md @@ -0,0 +1,54 @@ +# Locally Uncensored + +- **URL:** https://github.com/purpledoubled/locally-uncensored +- **License:** AGPL-3.0 +- **Status:** active (376 stars) +- **Last reviewed:** 2026-05-09 +- **Tags:** #ai #tooling + +## What it is + +Standalone desktop app (Tauri v2 + React 19 + Tailwind 4 + Rust) for running AI workloads — chat, image generation, video generation, coding agent — **entirely on the user's machine** with no cloud dependency or telemetry. Supports 20+ AI providers including Ollama, LM Studio, vLLM, KoboldCpp, llama.cpp, LocalAI, Jan, plus optional cloud (OpenAI, Anthropic) when wanted. + +Use case is "local-first AI" — privacy, offline use, no API bills, no content moderation gates. + +## When to install + +- Privacy-sensitive workflows (legal review, medical drafts, sensitive research) where cloud LLMs are off-limits. +- Offline development environments (air-gapped, on-plane, restricted networks). +- Want one desktop app that unifies chat / image / video / code agent across multiple local backends. +- Studying a clean Tauri v2 + React 19 reference implementation. + +## When NOT to install + +- No local GPU or modest hardware — local image/video gen needs serious VRAM. +- Need multi-user / shared deployment — this is desktop-single-user. +- AGPL-3.0 incompatible with your distribution plans. +- "Uncensored" is a compliance concern in your jurisdiction or vertical (regulated industries, public sector). + +## How to install + +Download the platform installer from upstream releases (Windows `.exe`, macOS, Linux) or build from source: + +```bash +git clone https://github.com/purpledoubled/locally-uncensored +cd locally-uncensored +pnpm install && pnpm tauri build +``` + +## Fit signals + +- Project requires offline / air-gapped LLM use. +- Team already runs Ollama, LM Studio, or vLLM locally. +- Stack uses Rust + Tauri or you're considering them — strong reference. + +## Conflicts and overlaps + +- Overlap with `openhuman` (also local-first personal agent) — OpenHuman is integration-heavy; Locally Uncensored is generation-heavy. +- Composes with any of the local LLM serving packs (it's a frontend, they're backends). + +## Notes + +- ⚠️ **AGPL-3.0** — same caveats as other AGPL packs. Personal/internal use is safe; redistribution as modified SaaS triggers the network-use clause. +- ⚠️ "Uncensored" framing — review what you're generating. The app removes provider-side moderation; you remain legally responsible for outputs in your jurisdiction. +- Tauri v2 + React 19 + Tailwind 4 is a clean reference stack. diff --git a/docs/registry/packs/mirothinker.md b/docs/registry/packs/mirothinker.md new file mode 100644 index 0000000..7178ece --- /dev/null +++ b/docs/registry/packs/mirothinker.md @@ -0,0 +1,55 @@ +# MiroThinker + +- **URL:** https://github.com/MiroMindAI/MiroThinker +- **License:** Apache-2.0 +- **Status:** active (8.2k stars) +- **Last reviewed:** 2026-05-09 +- **Tags:** #ai #agents #research + +## What it is + +A deep-research agent optimized for long-horizon, multi-step reasoning tasks (web browsing, code execution, information synthesis across domains). Ships open-source agent models built on Qwen plus a complete agent framework (MiroFlow) with benchmarking infrastructure. State-of-the-art on multiple research benchmarks. + +Author: MiroMindAI (the org). **Distinct from `miroshark`** (different author, different problem — MiroShark simulates synthetic crowds; MiroThinker is a research agent). + +## When to install + +- Need an open-source alternative to closed deep-research agents (Perplexity, GPT Deep Research) that you can self-host. +- Long-horizon research tasks — literature review, market research, multi-source synthesis with sources. +- Want to fine-tune or extend a research-agent model with your own data. +- Benchmarking research-agent quality against open baselines. + +## When NOT to install + +- Single-question Q&A — overkill, use a standard chat model with retrieval. +- Time-to-first-token sensitive UX — deep-research agents are slow by design. +- No GPU infrastructure available — serving Qwen-based models needs SGLang/vLLM and decent GPUs. + +## How to install + +```bash +git clone https://github.com/MiroMindAI/MiroThinker +cd MiroThinker +# Follow upstream README for SGLang/vLLM serving + MCP tool wiring +``` + +Tools are integrated via MCP, so it composes with existing MCP servers in the AI Dev OS ecosystem. + +## Fit signals + +- Project needs autonomous multi-step research (Sprint 1+ of `market-research-agent` or `research-waves` skill). +- Team is comfortable hosting open-source LLM stacks. +- Need a research benchmark / baseline for an internal research agent. + +## Conflicts and overlaps + +- Overlaps with `auto-research-claw` (paper-output research pipeline) — AutoResearchClaw is more focused on producing finished papers; MiroThinker is more general research-agent infra. +- Overlaps with `feynman-is` (external resource — local AI research agent live tool) — Feynman is a hosted product, MiroThinker is self-host code. +- Composes with `pathway-llm-app` (real-time RAG) for grounding. + +## Notes + +- Apache-2.0 — safe for commercial use with attribution. +- Built on Qwen base models — review Qwen license terms separately if redistributing weights. +- 8.2k stars at last review — active community, fast-moving research surface. +- Use MCP integration as the primary extension point rather than forking. diff --git a/docs/registry/packs/openhuman.md b/docs/registry/packs/openhuman.md new file mode 100644 index 0000000..cd320fe --- /dev/null +++ b/docs/registry/packs/openhuman.md @@ -0,0 +1,56 @@ +# OpenHuman + +- **URL:** https://github.com/tinyhumansai/openhuman +- **License:** GPL-3.0 +- **Status:** active beta (506 stars) +- **Last reviewed:** 2026-05-09 +- **Tags:** #agents #productivity #tooling + +## What it is + +A "personal AI super intelligence" — an open-source agentic assistant that integrates into daily workflows. Ships with **118+ third-party integrations** (Gmail, Notion, GitHub, Slack, Calendar, etc.), persistent memory management ("Memory Tree"), token-compression ("TokenJuice"), multi-channel communication, and supports local AI via Ollama or cloud providers. + +Built primarily in Rust + TypeScript. Runs as a desktop + CLI app. + +## When to install + +- Want a self-hosted alternative to closed personal assistants (Claude Desktop, ChatGPT Tasks, Pi). +- Need broad integration coverage across personal tools without writing each connector by hand. +- Privacy-sensitive personal use — Ollama backend keeps everything local. +- Building a "second brain" with persistent memory across sessions and tools. + +## When NOT to install + +- Production multi-user SaaS — this is a personal-use harness, not a multi-tenant platform. +- GPL-3.0 is a problem for your distribution model — review carefully if bundling. +- You only need 1-2 integrations — overkill, build them directly. +- Beta-tier reliability is unacceptable. + +## How to install + +```bash +git clone https://github.com/tinyhumansai/openhuman +cd openhuman +# Follow upstream — Rust toolchain + npm for the TS workspace +``` + +Desktop installers may also be published; check upstream releases. + +## Fit signals + +- Personal productivity automation is in scope. +- Team is comfortable running Rust binaries locally. +- Need multi-channel comms (email + chat + tasks) routed through one agent. +- Already using Ollama or local LLM stacks. + +## Conflicts and overlaps + +- Heavy overlap with `optio` (K8s ticket-to-PR workflow) — different focus (Optio = team ops, OpenHuman = personal ops). +- Overlap with `agentic-inbox` (Cloudflare-based email triage) — agentic-inbox is server-side / scoped to inbox; OpenHuman is desktop / general-purpose. +- Composes with `everything-claude-code:chief-of-staff` agent (multi-channel triage) and `everything-claude-code:email-ops` skill. + +## Notes + +- ⚠️ **GPL-3.0** — strong copyleft. Forking and redistributing modifications requires same license. Personal use is fine; commercial bundling needs careful review. +- Beta — APIs and integrations may shift. Track upstream releases. +- Memory Tree + TokenJuice are interesting primitives for any agent harness — even if you don't adopt OpenHuman wholesale, study these patterns. diff --git a/docs/registry/packs/quran-database.md b/docs/registry/packs/quran-database.md new file mode 100644 index 0000000..d692c3e --- /dev/null +++ b/docs/registry/packs/quran-database.md @@ -0,0 +1,54 @@ +# Quran Database + +- **URL:** https://github.com/AbdullahGhanem/quran-database +- **License:** check upstream (Quran text is public domain; software license not explicitly stated) +- **Status:** active (1.7k stars) +- **Last reviewed:** 2026-05-09 +- **Tags:** #learning #research #dataset + +## What it is + +A comprehensive **MySQL database dump** of the complete Quran text with multiple translations and editions. Distributed as a ~187 MB ZIP containing a single SQL dump. Provides structured, queryable access to Quranic content (verses, surahs, translations, scholarly metadata). + +Useful as a foundation dataset for any application that needs Quranic text — religious-tech apps, NLP / RAG pipelines on Arabic religious corpora, study tools, multilingual lookups. + +## When to install + +- Building a Quran-related application (study tool, audio sync, search, daily verse, etc.). +- Need a structured Quran dataset to ground a RAG / LLM agent. +- Comparing translations across editions programmatically. +- Researching Arabic NLP, transliteration, or theological text analysis. + +## When NOT to install + +- Need only one or two verses — fetch from a public API instead. +- Stack does not use MySQL (and you don't want to convert the dump). +- Religious-content distribution constraints in your jurisdiction or product context — review before shipping. + +## How to install + +```bash +# Download the ZIP from upstream releases +unzip quran-database.zip +mysql -u user -p quran < quran-database.sql +``` + +To convert to PostgreSQL / SQLite, use a standard MySQL-to-X migration tool. + +## Fit signals + +- Religious-tech, Islamic-edu, or Arabic-NLP project. +- Need an offline / self-hosted Quranic dataset (no API rate limits). +- Building a multilingual lookup that needs multiple translation editions side by side. + +## Conflicts and overlaps + +- No direct overlap with other packs in the registry — this is the first religious-text dataset cataloged. +- Composes with `crawlee` (if augmenting with tafsir/exegesis from public sites) and `pathway-llm-app` (if building a RAG over the corpus). + +## Notes + +- ✅ **Canonical Quran text is widely treated as public domain** in many contexts, but legal interpretation can vary by jurisdiction and distribution context. +- ⚠️ **Software license not explicitly stated** in the repo at last review — check the repository's LICENSE file or contact maintainers before redistributing the schema or scripts as part of a commercial product. +- ⚠️ Translation editions may carry their own copyright/license terms — verify per edition before bundling or redistribution. +- 1.7k stars — well-known reference in the Islamic-tech community. diff --git a/docs/registry/packs/ruflo.md b/docs/registry/packs/ruflo.md index 5c95a34..644abf0 100644 --- a/docs/registry/packs/ruflo.md +++ b/docs/registry/packs/ruflo.md @@ -2,8 +2,8 @@ - **URL:** https://github.com/ruvnet/ruflo - **License:** MIT -- **Status:** active (v3.6.10 released 2026-04-30, 6,117 commits, 34.1k stars) -- **Last reviewed:** 2026-04-30 +- **Status:** active (v3.6.30 released 2026-05-05, 6,376 commits, 47.7k stars) +- **Last reviewed:** 2026-05-09 - **Tags:** #ecosystem #agents #tooling ## What it is diff --git a/docs/registry/packs/understand-anything.md b/docs/registry/packs/understand-anything.md new file mode 100644 index 0000000..8f72635 --- /dev/null +++ b/docs/registry/packs/understand-anything.md @@ -0,0 +1,52 @@ +# Understand-Anything + +- **URL:** https://github.com/lum1104/understand-anything +- **License:** MIT +- **Status:** active (13.7k stars) +- **Last reviewed:** 2026-05-09 +- **Tags:** #tooling #productivity #ai #ecosystem + +## What it is + +Multi-platform AI plugin that turns any codebase, knowledge base, or docs collection into an **interactive knowledge graph** — files, functions, classes, and dependencies become an explorable, queryable graph. Compatible with Claude Code, Cursor, VS Code Copilot, Copilot CLI, Codex, Gemini CLI, and others through a specialized multi-agent analysis pipeline. + +Especially useful for codebase onboarding, business-logic flow mapping, and understanding unfamiliar projects quickly. + +## When to install + +- Joining a large or unfamiliar codebase — accelerates onboarding by visualizing structure. +- Auditing a project's surface area before refactor or major feature work. +- Need to explain a codebase to a non-dev stakeholder — graph visualization beats text walls. +- Cross-tool environment (your team uses Claude Code + Cursor + VS Code) — same plugin works across all. + +## When NOT to install + +- Tiny single-file projects — graph adds no signal over `ls`. +- Highly proprietary codebases where you cannot route source through external services (review network policy). +- Pure prose / non-code knowledge bases where AST graphs add no value. + +## How to install + +Multi-platform plugin — install pattern depends on host (Claude Code plugin manifest, Cursor extension, VS Code extension, etc.). See upstream README per platform. + +Visit: https://github.com/lum1104/understand-anything + +## Fit signals + +- Project is medium-to-large (hundreds of files, multiple domains). +- Onboarding new contributors is a recurring activity. +- Team uses multiple AI coding agents and wants a shared mental model. +- Need to visually explain dependencies for architecture review. + +## Conflicts and overlaps + +- Overlaps with `everything-claude-code:codebase-onboarding` skill — UAA gives the visual graph, the skill gives the structured doc. Complementary. +- Overlaps with `everything-claude-code:code-tour` skill — both navigate code, different UX (tour vs graph). +- Composes well with `claude-mem:smart-explore` for token-cheap structural search. + +## Notes + +- MIT — safe for commercial use. +- 13.7k stars — strong community signal. +- Multi-platform plugin architecture means you can adopt it without locking in to a single AI host. +- Review network/data policy if codebase routes through any external service. diff --git a/docs/registry/tags/agents-marketplace.md b/docs/registry/tags/agents-marketplace.md index 4cf6b5c..9976350 100644 --- a/docs/registry/tags/agents-marketplace.md +++ b/docs/registry/tags/agents-marketplace.md @@ -17,6 +17,7 @@ These differ from `#foundations` (canonical / official) and from `#ecosystem` (b | [`optio`](../packs/optio.md) | K8s-based ticket-to-PR workflow orchestration | community | | [`open-design`](../packs/open-design.md) | Local-first AI design generation across 10 agent CLIs | community | | [`trailofbits-skills`](../packs/trailofbits-skills.md) | Security-review skill bundle | Trail of Bits | +| [`googleworkspace-cli`](../packs/googleworkspace-cli.md) | ⭐ 100+ SKILL.md files for Gmail / Drive / Docs / Calendar / Sheets | Google (official) | ## When this tag is relevant diff --git a/docs/registry/tags/ai.md b/docs/registry/tags/ai.md index fe672fe..feb7888 100644 --- a/docs/registry/tags/ai.md +++ b/docs/registry/tags/ai.md @@ -26,6 +26,10 @@ Packs related to generative AI, LLMs, agents, and AI-powered applications. | [`kronos`](../packs/kronos.md) | Foundation model for financial candlesticks (AAAI 2026) | | [`diagram-design`](../packs/diagram-design.md) | Editorial-quality diagrams (Claude Code skill) | | [`miroshark`](../packs/miroshark.md) | Swarm intelligence simulation (synthetic crowd reactions) | +| [`mirothinker`](../packs/mirothinker.md) | Open-source deep-research agent (Qwen + MCP) | +| [`understand-anything`](../packs/understand-anything.md) | Codebase-to-knowledge-graph multi-platform plugin | +| [`crawlee`](../packs/crawlee.md) | Web scraping for LLM/RAG/GPT data ingestion (Apify) | +| [`locally-uncensored`](../packs/locally-uncensored.md) | Local-first chat / image / video / coding agent (desktop) | ## When this tag is relevant diff --git a/docs/registry/tags/data-engineering.md b/docs/registry/tags/data-engineering.md index 3fa3688..c0625dc 100644 --- a/docs/registry/tags/data-engineering.md +++ b/docs/registry/tags/data-engineering.md @@ -8,6 +8,7 @@ Packs related to data pipelines, warehousing, orchestration, ETL, and data-engin | [`data-engineering-zoomcamp`](../packs/data-engineering-zoomcamp.md) | DE bootcamp (Docker, Spark, Kafka, dbt) | | [`pathway-llm-app`](../packs/pathway-llm-app.md) | Real-time RAG with live data sync | | [`microsoft-data-science-for-beginners`](../packs/microsoft-data-science-for-beginners.md) | DS curriculum (touches DE basics) | +| [`crawlee`](../packs/crawlee.md) | Production web scraping for LLM/RAG ingestion (Apify) | ## When this tag is relevant diff --git a/docs/registry/tags/design.md b/docs/registry/tags/design.md index 94d9c1d..d81c6af 100644 --- a/docs/registry/tags/design.md +++ b/docs/registry/tags/design.md @@ -14,6 +14,8 @@ Packs in this category address UI/UX exploration, design systems, and visual pro | [`app-store-screenshots`](../packs/app-store-screenshots.md) | App Store + Play Store screenshot generator | | [`diagram-design`](../packs/diagram-design.md) | Editorial-quality diagrams (14 types) | | [`phantom-ui`](../packs/phantom-ui.md) | Auto-generated skeleton loading screens (Web Component) | +| [`isomiddleearth`](../packs/isomiddleearth.md) | Reference Next.js 16 / React 19 / Zustand isometric editor | +| [`ascii-draw`](../packs/ascii-draw.md) | Desktop ASCII diagram / flowchart drawing tool (GTK4) | ## External resources (not installable) diff --git a/docs/registry/tags/ecosystem.md b/docs/registry/tags/ecosystem.md index 87e1a22..bfbc6de 100644 --- a/docs/registry/tags/ecosystem.md +++ b/docs/registry/tags/ecosystem.md @@ -13,6 +13,8 @@ Broad, opinionated expansions of the Claude Code / agentic workflow ecosystem. T | [`optio`](../packs/optio.md) | K8s-based ticket-to-PR workflow orchestration | | [`vibeyard`](../packs/vibeyard.md) | Electron IDE for managing parallel AI agent sessions | | [`slavingia-skills`](../packs/slavingia-skills.md) | ⭐ Minimalist Entrepreneur 10-skill plugin (Sahil Lavingia) | +| [`understand-anything`](../packs/understand-anything.md) | Multi-platform plugin for codebase-as-knowledge-graph | +| [`claude-code-harness`](../packs/claude-code-harness.md) | Plan→Work→Review→Release harness with Go-native guardrails | ## When this tag is relevant diff --git a/docs/registry/tags/learning.md b/docs/registry/tags/learning.md index c0654df..0809f92 100644 --- a/docs/registry/tags/learning.md +++ b/docs/registry/tags/learning.md @@ -18,6 +18,7 @@ Curated learning curricula, courses, books, and reference handbooks. These are s | [`data-engineering-zoomcamp`](../packs/data-engineering-zoomcamp.md) | Free 9-week DE course | | [`microsoft-data-science-for-beginners`](../packs/microsoft-data-science-for-beginners.md) | Data science fundamentals | | [`slavingia-skills`](../packs/slavingia-skills.md) | ⭐ Minimalist Entrepreneur 10-skill business playbook | +| [`quran-database`](../packs/quran-database.md) | MySQL Quran dataset (text + multiple translations; license: check upstream) | ## When this tag is relevant diff --git a/docs/registry/tags/productivity.md b/docs/registry/tags/productivity.md index 360de57..3b96ccb 100644 --- a/docs/registry/tags/productivity.md +++ b/docs/registry/tags/productivity.md @@ -7,6 +7,10 @@ Packs and resources for personal-productivity products: email assistants, resear | [`agentic-inbox`](../packs/agentic-inbox.md) | pack | Self-hosted AI email client (Cloudflare Workers) | | [`feynman-is`](../external-resources/feynman-is.md) | external | Local AI research agent | | [`vibeyard`](../packs/vibeyard.md) | pack | Electron IDE for parallel AI agent sessions | +| [`understand-anything`](../packs/understand-anything.md) | pack | Codebase-as-knowledge-graph onboarding plugin | +| [`openhuman`](../packs/openhuman.md) | pack | Personal AI agent (118+ integrations, local-first) | +| [`cryptpad`](../packs/cryptpad.md) | pack | E2E-encrypted collaborative office suite (privacy-first) | +| [`googleworkspace-cli`](../packs/googleworkspace-cli.md) | pack | Official Google Workspace CLI for human + agent use | ## When this tag is relevant diff --git a/docs/registry/tags/research.md b/docs/registry/tags/research.md index df7a98d..f36d59a 100644 --- a/docs/registry/tags/research.md +++ b/docs/registry/tags/research.md @@ -11,6 +11,8 @@ Packs and resources for research-oriented work: replicating papers, literature r | [`kronos`](../packs/kronos.md) | pack | Foundation model for financial candlesticks (AAAI 2026) | | [`miroshark`](../packs/miroshark.md) | pack | Swarm intelligence simulation engine | | [`sci-bot`](../external-resources/sci-bot.md) | external | AI research assistant grounded in Sci-Hub papers | +| [`mirothinker`](../packs/mirothinker.md) | pack | Open-source deep-research agent (Qwen + MCP) | +| [`quran-database`](../packs/quran-database.md) | pack | Quranic text dataset for religious-text NLP / RAG | ## When this tag is relevant diff --git a/docs/registry/tags/security.md b/docs/registry/tags/security.md index a0dfb1e..71d2ca9 100644 --- a/docs/registry/tags/security.md +++ b/docs/registry/tags/security.md @@ -9,6 +9,7 @@ Packs in this category address application security, secret scanning, hardening, | [`cis-hardening-guide`](../packs/cis-hardening-guide.md) | Infrastructure hardening reference | | [`yara-style-guide`](../packs/yara-style-guide.md) | Malware detection rule style | | [`st3gg`](../packs/st3gg.md) | Steganography suite (red-team + blue-team) | +| [`cryptpad`](../packs/cryptpad.md) | E2E-encrypted collaborative office suite (zero-trust host) | ## When this tag is relevant diff --git a/session-log/2026-05-09-registry-additions.md b/session-log/2026-05-09-registry-additions.md new file mode 100644 index 0000000..9f0ed1b --- /dev/null +++ b/session-log/2026-05-09-registry-additions.md @@ -0,0 +1,75 @@ +# 2026-05-09 — Registry additions (14 new packs + 3 refreshes) + +## Why this session + +User dropped a list of 17 GitHub repos discovered through social-media browsing and asked for them to be analyzed and added to `docs/registry/` with proper documentation, categorized to fit cleanly with existing entries. Default workflow chosen: **rapid analysis (1-pager per repo)** + **dedicated branch** (`chore/registry-additions-2026-05-09` from `main`, leaving Sprint 0 PR #9 untouched). + +This is exactly the workflow the registry was designed for — see `feat/external-repo-registry-v0.4.0` rationale (April 2026): "Phase 2: continuous stream of newly discovered repos with minimal friction." + +## What was done + +### New packs (14) + +All written under `docs/registry/packs/` using the standard one-pager template: + +| Slug | Why it's interesting | +|------|----------------------| +| `mirothinker` | Open-source deep-research agent (Qwen + MCP), 8.2k stars | +| `understand-anything` | Multi-platform plugin turning codebases into queryable knowledge graphs, 13.7k stars | +| `agent-skills-eval` | Empirical A/B test runner for AI agent skills (judge-model grading) | +| `isomiddleearth` | Reference Next.js 16 / React 19 / Zustand isometric editor | +| `crawlee` | Production web scraping for LLM/RAG ingestion (Apify), 23.1k stars, Apache-2.0 | +| `openhuman` | Personal AI agent with 118+ integrations, Rust + TS, beta | +| `locally-uncensored` | Local-first chat / image / video / coding agent, Tauri v2 + React 19 | +| `agent-exchange` | Marketplace infrastructure for AI agents (10 Go services, ad-tech-style auctions) | +| `claude-code-harness` | Plan→Work→Review→Release harness for Claude Code, Go-native v4 "Hokage" | +| `jeweledtech-agentic-framework` | Multi-department business orchestration (15+ agents, 87 n8n workflows) | +| `quran-database` | MySQL Quran dataset (text + multiple translations; license pending upstream verification) | +| `cryptpad` | E2E-encrypted collaborative office suite (privacy-first), AGPL-3.0 | +| `googleworkspace-cli` | **Official** Google Workspace CLI (Rust), bundles 100+ SKILL.md files | +| `ascii-draw` | Desktop ASCII diagram drawing tool (GTK4 + Python) | + +### Refreshed (3 already existed in registry) + +- `fincept-terminal` → Stars 18.3k → 20.5k, last reviewed bumped +- `auto-research-claw` → Stars 11.8k → 12k, last reviewed bumped +- `ruflo` → Stars 34.1k → 47.7k, version 3.6.10 → 3.6.30, last reviewed bumped + +### INDEX and tag wiring + +- `docs/registry/INDEX.md`: 14 new rows added in alphabetical position; 3 existing rows refreshed; "Last index update" line bumped to 2026-05-09. +- `docs/registry/tags/ai.md` — added 4 packs (mirothinker, understand-anything, crawlee, locally-uncensored). +- `docs/registry/tags/agents-marketplace.md` — added googleworkspace-cli (qualifies as bundled SKILL.md collection). +- `docs/registry/tags/data-engineering.md` — added crawlee. +- `docs/registry/tags/design.md` — added isomiddleearth, ascii-draw. +- `docs/registry/tags/ecosystem.md` — added understand-anything, claude-code-harness. +- `docs/registry/tags/learning.md` — added quran-database. +- `docs/registry/tags/productivity.md` — added 4 packs (understand-anything, openhuman, cryptpad, googleworkspace-cli). +- `docs/registry/tags/research.md` — added mirothinker, quran-database. +- `docs/registry/tags/security.md` — added cryptpad. + +## Decisions worth preserving + +1. **MiroThinker is NOT MiroShark.** Same family of "Miro"-prefixed projects but different orgs (`MiroMindAI` vs `aaronjmars`) and different problems (research agent vs synthetic-crowd simulation). Both kept as separate packs with cross-references in their respective Notes sections. + +2. **`googleworkspace-cli` got the `agents-marketplace` tag** even though it's primarily a CLI tool. Reason: ships with **100+ bundled SKILL.md files** (40+ for Gmail/Drive/Docs/Calendar/Sheets alone), which qualifies under the agents-marketplace definition ("opinionated, curated, delivered as a discrete bundle, often with a single author/org signal"). Google as the official maintainer is a strong author signal. + +3. **`isomiddleearth` license is unstated upstream** — added the pack with explicit "view-only reference until upstream clarifies" warning. Tolkien IP is also flagged in Notes. + +4. **`quran-database` introduces `#dataset` tag** as metadata-only (no `tags/dataset.md` file yet). Other packs use ad-hoc tags like `#tooling`, `#infra`, `#experimental`, `#malware`, `#origin` without dedicated tag files — same pattern. If the registry accumulates more datasets, promote `#dataset` to a proper tag file. + +5. **Copyleft-licensed packs flagged consistently** — `cryptpad` (AGPL-3.0), `locally-uncensored` (AGPL-3.0), `openhuman` (GPL-3.0), `fincept-terminal` (AGPL+commercial) all carry explicit license warnings in their Notes. Vibe-coders may bundle without realizing — the warnings are deliberate. + +6. **Refresh-only for the 3 already-cataloged repos** — bumped "Last reviewed" + version/star numbers in pack file and INDEX. Did not rewrite content; the original analyses remain valid. + +## Branch / PR plan + +- Branch: `chore/registry-additions-2026-05-09` from `main`. +- Single commit covering all additions + refreshes + index/tag wiring. +- PR target: `main`. Independent of Sprint 0 PR #9 (no overlap). + +## Skipped / open items + +- **No new tag files created** for `#dataset`, `#tooling`, `#agents`, `#infra`, `#experimental`, `#ecosystem` (some already exist as files, some don't). Current ad-hoc pattern preserved. +- **Did not run `os-self-test`** — could be done after merge to validate cross-references are intact. +- **No update to `WIZARD.md` or `registry-pick` skill** — none needed since the new packs are discoverable through INDEX + tags via the existing surface. diff --git a/session-log/INDEX.md b/session-log/INDEX.md index cc15602..ebc78e4 100644 --- a/session-log/INDEX.md +++ b/session-log/INDEX.md @@ -6,6 +6,7 @@ Each entry captures the reasoning behind a material decision: what was true befo | Date | Topic | File | |---|---|---| +| 2026-05-09 | Registry additions — 14 new packs cataloged, 3 refreshed (Phase 2 cadence) | [`2026-05-09-registry-additions.md`](2026-05-09-registry-additions.md) | | 2026-05-04 | v0.4.5 course vertical — workshop em 3 aulas (Experience Learning Perestroika) + 2 sistemas planejados + easter egg `[STAGE:X]` | [`2026-05-04-v0.4.5-course-vertical.md`](2026-05-04-v0.4.5-course-vertical.md) | | 2026-05-04 | v0.4.4 repo hardening — branch protection, GitHub Template, WIZARD Stage 0.5 (detach from OS-origin), CONTRIBUTING flow | [`2026-05-04-v0.4.4-repo-hardening.md`](2026-05-04-v0.4.4-repo-hardening.md) | | 2026-05-01 | v0.4.3 quick wins — ETHOS manifesto, /processize, /grow-sustainably, /multi-ai-review, agents-marketplace tag | [`2026-05-01-v0.4.3-quick-wins.md`](2026-05-01-v0.4.3-quick-wins.md) |