leo-lilinxiao/codex-autoresearch vs uditgoenka/autoresearch
Compare leo-lilinxiao/codex-autoresearch and uditgoenka/autoresearch using the same repository snapshot.
Related repositories are selected from shared categories, topics and mentions in the same lists. Order follows the export’s relevance signals. These links are independent of sponsorship and do not change the leaderboard.
leo-lilinxiao/codex-autoresearch
Run repeated, measured Git experiments toward a numeric target; keep improvements and revert failures. Use for autonomous optimization or managing an autoresearch run, not one-shot edits.
- GitHub stars
- 3,145
- Star growth over 7 days
- +1.8%
- Heat
- 6.0
- Before you install
- Network or broad shell
- Median activation tokens (estimate)
- 562 tokens
- Project activity
- Active
- License
- MIT
- Supported agents
- Not available
uditgoenka/autoresearch
Claude Autoresearch Skill — Autonomous goal-directed iteration for Claude Code. Inspired by Karpathy's autoresearch. Modify → Verify → Keep/Discard → Repeat forever.
- GitHub stars
- 6,811
- Star growth over 7 days
- +0.4%
- Heat
- 4.4
- Before you install
- Network or broad shell
- Median activation tokens (estimate)
- 2,456 tokens
- Project activity
- Active
- License
- MIT
- Supported agents
- Claude Code, Codex, OpenCode
Token cost is the median frontmatter plus SKILL.md estimate per skill. Reference files load on demand and scripts run separately. The estimate is unavailable when skill records or token counts are incomplete. Missing growth or heat history stays unmeasured, never zero.
Automated signals, not a guarantee Methodology
