AI Repo Acceleration Report

Which AI repos on GitHub are speeding up, and which have gone flat — measured from our own daily star snapshots. Updated 2026-08-08.

This week in numbers

whatstrending.ai stores its own daily star snapshot for 954 AI repositories on GitHub. As of 2026-08-08, 172 of them have a snapshot from the last 6 days, and 67 have enough history for a current pace.

The median currently tracked AI repo is gaining 203 stars a day.

mattpocock/skills is the fastest-growing AI repo we track right now at 1,569 stars a day (+10,983 stars between 2026-08-01 and 2026-08-08).

addyosmani/agent-skills accelerated hardest: from 133 to 390 stars a day, a jump of 257 stars a day.

paperswithbacktest/awesome-systematic-trading cooled hardest: from 614 down to 118 stars a day.

Of the 35 repos where we can measure both a current and a prior window, 5 are accelerating, 26 are steady, 4 have stalled and 0 are losing stars.

Tracked repos9546,104 snapshots on 51 days
History starts2026-06-19nothing stored from before this date
Snapshotted recently172in the last 6 days
Measurable pace67have two snapshots 5-12 days apart
Median pace203stars per day, across those 67
Stars gained+135,696total across those 67 repos
Accelerating5of 35 with two windows
Steady26of 35 with two windows
Stalled4of 35 with two windows
Declining0of 35 with two windows

Accelerating hardest

Biggest increase in stars per day versus each repo’s own prior window — not versus other repos. Both windows and their exact dates are shown, because a 5-day and a 10-day window are not the same measurement.

  1. 1
    addyosmani/agent-skills
    133/day 390/day +257/day JavaScript
    now: 2026-08-01 → 2026-08-08 (7 days, +2,730 stars) · before: 2026-07-27 → 2026-08-01 (5 days, +667 stars)
  2. 2
    Panniantong/Agent-Reach
    399/day 639/day +240/day Python
    now: 2026-08-01 → 2026-08-06 (5 days, +3,196 stars) · before: 2026-07-25 → 2026-08-01 (7 days, +2,795 stars)
  3. 3
    mvanhorn/last30days-skill
    238/day 458/day +220/day Python
    now: 2026-07-29 → 2026-08-03 (5 days, +2,289 stars) · before: 2026-07-24 → 2026-07-29 (5 days, +1,189 stars)
  4. 4
    obra/superpowers
    493/day 608/day +114/day Shell
    now: 2026-07-30 → 2026-08-08 (9 days, +5,468 stars) · before: 2026-07-25 → 2026-07-30 (5 days, +2,466 stars)
  5. 5
    firecrawl/firecrawl
    437/day 537/day +100/day TypeScript
    now: 2026-07-29 → 2026-08-08 (10 days, +5,374 stars) · before: 2026-07-24 → 2026-07-29 (5 days, +2,186 stars)
  6. 6
    CherryHQ/cherry-studio
    39/day 115/day +76/day TypeScript
    now: 2026-08-01 → 2026-08-07 (6 days, +687 stars) · before: 2026-07-27 → 2026-08-01 (5 days, +193 stars)
  7. 7
    NousResearch/hermes-agent
    485/day 550/day +65/day Python
    now: 2026-08-02 → 2026-08-07 (5 days, +2,752 stars) · before: 2026-07-28 → 2026-08-02 (5 days, +2,425 stars)
  8. 8
    K-Dense-AI/scientific-agent-skills
    87/day 93/day +6.5/day Python
    now: 2026-08-01 → 2026-08-07 (6 days, +560 stars) · before: 2026-07-27 → 2026-08-01 (5 days, +434 stars)
  9. 9
    iOfficeAI/AionUi
    64/day 66/day +2.4/day TypeScript
    now: 2026-08-01 → 2026-08-08 (7 days, +465 stars) · before: 2026-07-27 → 2026-08-01 (5 days, +320 stars)
  10. 10
    KnockOutEZ/wigolo
    60/day 63/day +2.3/day TypeScript
    now: 2026-07-28 → 2026-08-07 (10 days, +627 stars) · before: 2026-07-23 → 2026-07-28 (5 days, +302 stars)

Stalled: were growing, now flat

Repos that were gaining at least 1 star a day in their prior window and have since dropped to near-flat — at or below a third of their own previous rate. Ordered by how many stars per day they lost.

  1. 1
    paperswithbacktest/awesome-systematic-trading
    614/day 118/day −496/day Python
    now: 2026-08-02 → 2026-08-08 (6 days, +709 stars) · before: 2026-07-28 → 2026-08-02 (5 days, +3,072 stars)
  2. 2
    ComposioHQ/awesome-claude-skills
    481/day 97/day −385/day Python
    now: 2026-07-27 → 2026-08-07 (11 days, +1,063 stars) · before: 2026-07-22 → 2026-07-27 (5 days, +2,406 stars)
  3. 3
    pascalorg/editor
    503/day 130/day −374/day TypeScript
    now: 2026-08-01 → 2026-08-06 (5 days, +648 stars) · before: 2026-07-27 → 2026-08-01 (5 days, +2,517 stars)
  4. 4
    moeru-ai/airi
    499/day 144/day −355/day TypeScript
    now: 2026-08-01 → 2026-08-06 (5 days, +718 stars) · before: 2026-07-27 → 2026-08-01 (5 days, +2,494 stars)

Fastest by absolute pace

Raw stars per day in the current window, regardless of whether that is faster or slower than before. This is the leaderboard; the section above is the movement.

  1. 1
    mattpocock/skills
    1,569/day +10,983 stars was 1,651/day Shell
    2026-08-01 → 2026-08-08 (7 days, +10,983 stars) · 208,783 stars at our 2026-08-08 snapshot
  2. 2
    zhaoxuya520/reverse-skill
    1,438/day +7,188 stars no prior window PowerShell
    2026-08-03 → 2026-08-08 (5 days, +7,188 stars) · 20,528 stars at our 2026-08-08 snapshot
  3. 3
    TencentCloud/TencentDB-Agent-Memory
    1,302/day +6,509 stars no prior window TypeScript
    2026-08-03 → 2026-08-08 (5 days, +6,509 stars) · 17,492 stars at our 2026-08-08 snapshot
  4. 4
    diegosouzapw/OmniRoute
    950/day +4,749 stars was 1,013/day TypeScript
    2026-07-29 → 2026-08-03 (5 days, +4,749 stars) · 37,876 stars at our 2026-08-03 snapshot
  5. 5
    lyogavin/airllm
    855/day +4,277 stars no prior window Jupyter Notebook
    2026-08-03 → 2026-08-08 (5 days, +4,277 stars) · 29,904 stars at our 2026-08-08 snapshot
  6. 6
    microsoft/AI-For-Beginners
    839/day +4,196 stars no prior window Jupyter Notebook
    2026-08-03 → 2026-08-08 (5 days, +4,196 stars) · 63,169 stars at our 2026-08-08 snapshot
  7. 7
    bojieli/ai-agent-book
    794/day +4,763 stars was 1,519/day Python
    2026-07-31 → 2026-08-06 (6 days, +4,763 stars) · 32,119 stars at our 2026-08-06 snapshot
  8. 8
    esengine/DeepSeek-Reasonix
    770/day +3,848 stars no prior window Go
    2026-08-03 → 2026-08-08 (5 days, +3,848 stars) · 32,889 stars at our 2026-08-08 snapshot
  9. 9
    block/buzz
    768/day +3,838 stars was 961/day Rust
    2026-08-03 → 2026-08-08 (5 days, +3,838 stars) · 24,935 stars at our 2026-08-08 snapshot
  10. 10
    Panniantong/Agent-Reach
    639/day +3,196 stars was 399/day Python
    2026-08-01 → 2026-08-06 (5 days, +3,196 stars) · 66,599 stars at our 2026-08-06 snapshot

What this measures — and what it does not

The data is ours. whatstrending.ai records its own star and fork snapshot for every repository in its trending index and keeps them. This page holds 6,104 snapshot rows across 954 repositories, recorded on 51 snapshot days. GitHub shows a repo’s star count today; it does not show whether that number is climbing faster or slower than it was a week ago. That is the only thing this page reports.

History starts 2026-06-19. The most recent snapshot in this report is 2026-08-08. We hold nothing from before 2026-06-19 and we do not reconstruct it. Days with no snapshot for a repo are simply absent — they are never interpolated.

How a window is picked. For each repo: its newest stored snapshot, then the newest snapshot 5–12 days earlier (the current window), then the newest snapshot 5–12 days before that (the prior window). Pace is star delta / whole days between the two snapshots. Acceleration is current pace − prior pace. A repo whose newest snapshot is more than 5 days behind the newest snapshot in the dataset is excluded entirely rather than reported with a stale window.

Why the counts are smaller than 954. We snapshot a repo while it is in the trending index, so coverage is dense for currently-trending repos and sparse for ones that have dropped out. Of 954 repos with any stored history, 172 have a snapshot from the last 6 days, 67 of those have a measurable current pace, and 35 also have a prior window to compare against. Every list on this page states which pool it is drawn from. Where a figure cannot be derived from stored rows, it is omitted rather than estimated.

Citing this. Free to quote with attribution to whatstrending.ai. The underlying repo list is available as JSON at /api/repos, and each repo’s full measured history is on its own page.

Which AI repos are accelerating fastest right now?

As of 2026-08-08, addyosmani/agent-skills shows the biggest jump in star growth rate among the repos whatstrending.ai snapshots: from 133 stars a day (2026-07-27 to 2026-08-01) to 390 stars a day (2026-08-01 to 2026-08-08), an increase of 257 stars a day. The full top 10 is on this page.

How do you measure whether an AI repo is accelerating or stalling?

whatstrending.ai stores its own daily star snapshot for every repo in its trending index (6104 snapshot rows across 954 repos on 51 snapshot days since 2026-06-19). For each repo we take its newest snapshot, the newest snapshot 5 to 12 days before that (the current window), and the newest snapshot 5 to 12 days before THAT (the prior window). Growth rate is the star delta divided by the whole days between the two snapshots. Acceleration is the current rate minus the prior rate. Nothing is estimated, interpolated or back-filled.

How many AI repos does whatstrending.ai track?

954 repositories have at least one stored star snapshot, recorded on 51 snapshot days since 2026-06-19. 172 of them have a snapshot from the last 6 days as of 2026-08-08, and 67 of those have enough history for a current growth rate.

All tracked repos → Sort by velocity → Accelerating repos → Trending AI Repo badge → Model usage trends →