Autonomous codebase improvement loop for Claude Code. Inspired by karpathy/autoresearch.
active 2026-03-26 → 2026-04-02 (UTC)
Activity over time
Daily event counts in the loaded window
Line chart, 8 days from 2026-03-26 to 2026-04-02. Pushes: 63 total, peak 17 in a day. Pull requests: 14 total, peak 6 in a day. Issues: 18 total, peak 6 in a day. Comments: 10 total, peak 5 in a day. Stars: 0 total, peak 0 in a day.
- Pushes
- Pull requests
- Issues
- Comments
- Stars
Stars, PRs, issues and forks are under-captured in the later part of this window. GH Archive progressively stopped capturing non-push events during 2026 — −95% or worse by the end of the window. Every series here except Pushes fades for that reason, so a decline above reflects the archive, not this repository. Pushes stay reliable throughout, so read them, and the contributor counts derived from them, as the real signal. Data health has the measurements.
Top contributors
Pushes, PRs, issues, reviews and comments — stars and forks excluded, so this is contribution rather than popularity
| Contributor | Contributions | Pushes | PRs | Comments |
|---|---|---|---|---|
| ipedro | 93 | 60 | 12 | 3 |
| Copilot | 12 | 2 | 2 | 7 |
| dependabot[bot] | 1 | 1 | 0 | 0 |
Recent activity
Latest issues, pull requests and releases
- Issue comment#87ipedro2026-04-02 09:26fix(idea-matrix): enforce parallel dispatch + no-tools constraint in agents
- Releaseipedro2026-04-02 03:39v0.4.0 — Mechanical AR Chain + Idea Matrix --brief
- Issue#87ipedro2026-04-02 02:20fix(idea-matrix): enforce parallel dispatch + no-tools constraint in agents
- Issue#85ipedro2026-04-02 00:15feat(idea-matrix): model escalation for synthesis step
- Issue#85ipedro2026-04-02 00:15feat(idea-matrix): model escalation for synthesis step
- Issue#84ipedro2026-04-02 00:15feat(idea-matrix): telemetry run folder (~/.autoimprove/runs/)
- Issue#84ipedro2026-04-02 00:15feat(idea-matrix): telemetry run folder (~/.autoimprove/runs/)
- Issue#84ipedro2026-04-02 00:15feat(idea-matrix): telemetry run folder (~/.autoimprove/runs/)
- Issue#38ipedro2026-04-01 21:57feat: Phase 4 — org-level grind loop on claudinho workflow metrics
- Issue#55ipedro2026-04-01 21:57spike: model routing optimization — find Goodhart-safe approach via idea-matrix
- Pull request#77ipedro2026-04-01 21:47
- Pull request#77ipedro2026-04-01 21:47
- Issue#76ipedro2026-04-01 21:12feat: Interactive experiment CRUD via Claude TUI (create/list/remove)
- Issue#52ipedro2026-04-01 20:49ar-retro: [LOW] harvest-themes.sh — filenames with newlines silently truncated
- Pull request#72ipedro2026-04-01 19:35
- Pull request#70Copilot2026-04-01 16:26
- Pull request#70Copilot2026-04-01 16:26
- Pull request#67ipedro2026-04-01 03:07
- Issue#56ipedro2026-03-30 08:09feat: cross-model calibration — use expensive model as benchmark to optimize cheap model prompts
- Issue#49ipedro2026-03-30 07:14[AR-sweep] Trust-tier incoherence after rollback — tier not reset
- Issue#40ipedro2026-03-29 15:08feat: add qualitative benchmark — prompt-testing score for skill/agent quality
- Pull request#28ipedro2026-03-29 01:14
- Pull request#27ipedro2026-03-29 00:47
- Pull request#27ipedro2026-03-29 00:46
- Issue#21ipedro2026-03-28 21:43spike: non-interactive autoimprove invocation — trigger cannot launch full loop via claude --print
Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 0 stars here means stars gained during the window, not the repo's star count.