中/EN
— reading now
2026.10.18Sunday·AI beyond the horizon·himiao.ai
Flash

GitHub Weekly

One issue a week, 5–8 projects that are actually worth your time. We run what we can and read the code of what we can’t — every pick says which. Plus a straight take: worth it or not, the catch, and who it’s for. Star counts are tracked by us, daily.

  1. Vol.1·Oct 17·Latest

    GitHub Weekly Vol.1: 8 Projects Worth Your Time

    1. graphify-labs/graphifyTestedAn agent skill turning any folder (code, docs, PDFs, images, audio/video) into a queryable knowledge graph — `/graphify .` — for 20+ agent tools.★ 127,596
    2. justvugg/colibriTestedA pure-C local LLM inference engine — 'tiny engine, immense model'. One hand-written C engine per model family; runs hundred-billion-parameter MoE models on plain CPU machines by keeping only the always-active part resident in RAM and paging experts from disk.★ 42,540
    3. alibaba/open-code-reviewTestedAlibaba's internal AI code-review CLI, now open source; a Go-based `ocr` tool that reads git diffs and reviews them with a configurable LLM, producing line-precise comments.★ 47,541
    4. mksglu/context-modeTestedAn MCP server fighting context-window waste for coding agents: sandboxed tools keep raw data out of context, SQLite + FTS5/BM25 provides session continuity across compactions, enforcing a 'LLM writes code to analyze, doesn't read data itself' paradigm across 17 clients.★ 30,460
    5. panniantong/agent-reachTestedAn internet capability layer for AI agents — picks, installs, and health-checks the most reliable access route per platform (X, Reddit, YouTube, Bilibili, Xiaohongshu, GitHub, RSS, 16+ channels).★ 98,612
    6. zhuyansen/new-term-radarTestedA daily radar for emerging AI terms: watches 7 signals (X accounts, Hugging Face trending, GitHub repo breakouts, AI-tool-site sitemaps…) to catch new model/product/repo names before keyword tools list them. Pure Python stdlib, plus 3 Agent Skills and an MCP server.★ 1,079
    7. cathrynlavery/diagram-designTestedAn agent skill that makes coding agents draw editorial-quality diagrams: ~45 diagram types, each emitted as one self-contained HTML+SVG file, with brand matching by reading your website. Tagline: 'No Mermaid slop.'★ 56,841
    8. Oldcircle/geo-sleuthReadAn open Agent Skill (SKILL.md + ~20 Python scripts) packaging photo-geolocation methodology into coding agents: hand the agent a photo and one sentence, and it runs scripted evidence-gathering to return coordinates, error radius, camera heading, and an evidence image.★ 3,203

    Read the issue →

How we pick and test

  1. 01We start from the week’s trending repos and check Hacker News and Reddit for real usage. Already-everywhere repos, marketing-only repos and bare demos are out.
  2. 02“Tested” means we installed and ran it; “Read” means we read the code and docs without running it. Every pick is labeled.
  3. 03Star counts and weekly growth come from our own daily readings of GitHub; no growth figure until we have 7 days of data.
  4. 04Takes are based only on our test results and public numbers. No paid placements.
Ask the cat