Menu

Not in the top 1,000 by GitHub stars

GAIR-NLP/ASI-Evolve

Engineering

History

Review weekly changes to SKILL.md before updating. Dates mark the weeks in which we observed each version; checks apply to the recorded commits.

View weekly changes

skills/evolve/SKILL.md

Week of Oct 26, 2026 · +0 / −0 lines

Before Network or broad shell → After Network or broad shell

210707a3d4a3 → 210707a3d4a3

Compare on GitHub

View on GitHubAdd to my skills

Verified commit 210707a3d4a3bb1cee458919462c07640d7dfff1

871GitHub stars
-1star in 7 days
1skill in this repository
1commit in 30 days

Heat

1.0
  • Star growth, last 7 daysNot available
  • Star growth, last 24 hoursNot available
  • Rank movement, last 7 daysNot available
  • Commits, last 30 daysNot available
  • Accounts asking about it through MCP, last 7 daysNot available
  • Hot votes this weekNot available

Heat never changes the order by stars, and sponsorship never touches it. How heat works

Stars over time

Aug 2, 2026Nov 6, 2026872

Individual skills in this repository

How it works

evolve

When your agent uses it

Run an ASI-Evolve style evaluator-driven search workflow for code, algorithms, prompts, or pipelines. Use when Codex needs to align the objective, scoring, evaluator, writable scope, and cognition first, then execute a preflight-gated learn/design/experiment/analyze loop with the bundled Evolve CLI instead of the repository's multi-agent pipeline.

What it loads into context

Frontmatter

80 tokens

Always available to the agent

SKILL.md

2,147 tokens

When the skill is activated

Reference documents

3,338 tokens

Loaded on request

Token counts are estimates. Scripts run separately; they are not counted as reference context.

The playbook

  1. Preserve the operating model
  2. Start with preflight
  3. Initialize cognition
  4. Run each evolution round explicitly
  5. Serialize database operations
  6. Make good round decisions
  7. Escalate back to the user when needed
  8. Avoid these failures
  9. Read when needed

Scripts executed separately

  • ../../main.py
  • scripts/evolve-brief
  • scripts/evolve-cognition
  • scripts/evolve-db
  • scripts/evolve-eval
  • scripts/evolve-files
  • scripts/evolve-summary
  • scripts/evolve_core/__init__.py
  • scripts/evolve_core/algorithms/__init__.py
  • scripts/evolve_core/algorithms/base.py
  • scripts/evolve_core/algorithms/factory.py
  • scripts/evolve_core/algorithms/greedy.py
  • scripts/evolve_core/algorithms/island.py
  • scripts/evolve_core/algorithms/random.py
  • scripts/evolve_core/algorithms/ucb1.py
  • scripts/evolve_core/best_snapshot.py
  • scripts/evolve_core/cli.py
  • scripts/evolve_core/cognition.py
  • scripts/evolve_core/database.py
  • scripts/evolve_core/diff.py
  • scripts/evolve_core/embedding.py
  • scripts/evolve_core/file_lock.py
  • scripts/evolve_core/run_state.py
  • scripts/evolve_core/sampling_config.py
  • scripts/evolve_core/structures.py
  • scripts/evolve_core/vector_index.py

Files

  • ../../main.pyruns2,516 bytes
  • SKILL.mdcore10,799 bytes
  • SKILL.mdcore10,799 bytes
  • SKILL.mdcore10,799 bytes
  • SKILL.mdcore10,799 bytes
  • SKILL.mdcore10,799 bytes
  • agents/openai.yamlasset316 bytes
  • references/architecture.mddoc1,519 bytes
  • references/operating_model.mddoc2,086 bytes
  • references/preflight.mddoc1,751 bytes
  • references/run_spec.mddoc1,935 bytes
  • references/toolbelt.mddoc7,899 bytes
  • scripts/evolve-briefruns344 bytes
  • scripts/evolve-cognitionruns330 bytes
  • scripts/evolve-dbruns322 bytes
  • scripts/evolve-evalruns326 bytes
  • scripts/evolve-filesruns327 bytes
  • scripts/evolve-summaryruns326 bytes
  • scripts/evolve_core/__init__.pyruns317 bytes
  • scripts/evolve_core/algorithms/__init__.pyruns385 bytes
  • scripts/evolve_core/algorithms/base.pyruns498 bytes
  • scripts/evolve_core/algorithms/factory.pyruns3,130 bytes
  • scripts/evolve_core/algorithms/greedy.pyruns415 bytes
  • scripts/evolve_core/algorithms/island.pyruns12,687 bytes
  • scripts/evolve_core/algorithms/random.pyruns409 bytes
  • scripts/evolve_core/algorithms/ucb1.pyruns1,716 bytes
  • scripts/evolve_core/best_snapshot.pyruns2,030 bytes
  • scripts/evolve_core/cli.pyruns28,095 bytes
  • scripts/evolve_core/cognition.pyruns4,656 bytes
  • scripts/evolve_core/database.pyruns8,143 bytes
  • scripts/evolve_core/diff.pyruns1,757 bytes
  • scripts/evolve_core/embedding.pyruns2,401 bytes
  • scripts/evolve_core/file_lock.pyruns1,239 bytes
  • scripts/evolve_core/run_state.pyruns11,949 bytes
  • scripts/evolve_core/sampling_config.pyruns6,281 bytes
  • scripts/evolve_core/structures.pyruns2,717 bytes
  • scripts/evolve_core/vector_index.pyruns6,961 bytes
Read SKILL.md

# Evolve

Treat this skill as the single-agent abstraction of `ASI-Evolve-dev`. Preserve the system's core idea: learn from prior knowledge, design the next candidate, run the evaluator, analyze the outcome, and feed the lesson back into the next round.

## Preserve the operating model
- Keep the four-stage loop intact: learn, design, experiment, analyze.
- Use the bundled wrappers in `scripts/` for stateful operations. Let the agent make orchestration decisions; do not recreate the repository's pipeline stack.
- The agent itself must choose mutations, run experiments, inspect outputs, and write round analyses step by step.
- Keep two memory systems alive for the whole run:
  - cognition store for approved external research, paper takeaways, and search-derived insights that may help future rounds
  - experiment database for every candidate, score, analysis, lineage, best snapshot, and lessons learned from actual experiments
- Treat `.evolve_runs/<run-name>/` as the source of truth for the active run.

## Start with preflight
View the complete source on GitHub

Skills included

  • evolve

Repository facts

License
Apache-2.0
Created
Mar 27, 2026
Last pushed
Oct 9, 2026
Forks
215
Open issues
0
Project activity
Active

Rank over the last 12 weeks

Now #999 · best #624 · down 188 in 30 days

  1. #977 on Nov 1
  2. #983 on Nov 2
  3. #989 on Nov 3
  4. #994 on Nov 4
  5. #995 on Nov 5
  6. #999 since Nov 6
39 earlier periods
  1. #616 on Aug 2
  2. #607 on Aug 8
  3. #617 on Aug 9
  4. #642 on Aug 10
  5. #650 on Aug 11
  6. #647 on Aug 12
  7. #644 on Aug 13
  8. #622 on Aug 15
  9. #663 on Aug 18
  10. #648 on Aug 20
  11. #653 on Aug 21
  12. #624 on Aug 22
  13. #677 on Aug 25
  14. #660 on Aug 26
  15. #669 on Aug 27
  16. #646 on Aug 29
  17. #679 on Sep 2
  18. #674 on Sep 6
  19. #708 on Sep 10
  20. #700 on Sep 13
  21. #713 on Sep 15
  22. #744 on Sep 22
  23. #711 on Sep 24
  24. #695 on Sep 25
  25. #773 on Sep 29
  26. #752 on Sep 30
  27. #734 on Oct 4
  28. #811 on Oct 8
  29. #791 on Oct 13
  30. #795 on Oct 15
  31. #778 on Oct 17
  32. #764 on Oct 18
  33. #810 on Oct 19
  34. #809 on Oct 22
  35. #817 on Oct 23
  36. #945 on Oct 25
  37. #967 on Oct 28
  38. #971 from Oct 29 to Oct 30
  39. #975 on Oct 31

Before you install

Network or broad shell

Automated signals, not a guarantee

Checked Nov 4, 2026

Not fully analyzed: some files were too long, unreadable or missing. This never lowers the level.

Show 4 notes
  • skills/evolve/SKILL.md:1 Analysis incomplete: Referenced executable target unavailable; not analyzed
  • skills/evolve/SKILL.md:1 Analysis incomplete: Referenced executable target unavailable; not analyzed
  • skills/evolve/SKILL.md:1 Analysis incomplete: Referenced executable target unavailable; not analyzed
  • skills/evolve/SKILL.md:1 Analysis incomplete: Referenced executable target unavailable; not analyzed
Show all 27 signals by file

main.py

  • Executable scriptline 1
    #!/usr/bin/env python3

skills/evolve/scripts/evolve-brief

  • Executable scriptline 1
    #!/usr/bin/env python3

skills/evolve/scripts/evolve-cognition

  • Executable scriptline 1
    #!/usr/bin/env python3

skills/evolve/scripts/evolve-db

  • Executable scriptline 1
    #!/usr/bin/env python3

skills/evolve/scripts/evolve-eval

  • Executable scriptline 1
    #!/usr/bin/env python3

skills/evolve/scripts/evolve-files

  • Executable scriptline 1
    #!/usr/bin/env python3

skills/evolve/scripts/evolve-summary

  • Executable scriptline 1
    #!/usr/bin/env python3

skills/evolve/scripts/evolve_core/__init__.py

  • Executable scriptline 1
    Executable file

skills/evolve/scripts/evolve_core/algorithms/__init__.py

  • Executable scriptline 1
    Executable file

skills/evolve/scripts/evolve_core/algorithms/base.py

  • Executable scriptline 1
    Executable file

skills/evolve/scripts/evolve_core/algorithms/factory.py

  • Executable scriptline 1
    Executable file

skills/evolve/scripts/evolve_core/algorithms/greedy.py

  • Executable scriptline 1
    Executable file

skills/evolve/scripts/evolve_core/algorithms/island.py

  • Executable scriptline 1
    Executable file

skills/evolve/scripts/evolve_core/algorithms/random.py

  • Executable scriptline 1
    Executable file

skills/evolve/scripts/evolve_core/algorithms/ucb1.py

  • Executable scriptline 1
    Executable file

skills/evolve/scripts/evolve_core/best_snapshot.py

  • Executable scriptline 1
    Executable file

skills/evolve/scripts/evolve_core/cli.py

  • Executable scriptline 1
    Executable file

skills/evolve/scripts/evolve_core/cognition.py

  • Executable scriptline 1
    Executable file

skills/evolve/scripts/evolve_core/database.py

  • Executable scriptline 1
    Executable file
  • Access to secretsline 212
    [Credential-related source omitted]

skills/evolve/scripts/evolve_core/diff.py

  • Executable scriptline 1
    Executable file

skills/evolve/scripts/evolve_core/embedding.py

  • Executable scriptline 1
    Executable file

skills/evolve/scripts/evolve_core/file_lock.py

  • Executable scriptline 1
    Executable file

skills/evolve/scripts/evolve_core/run_state.py

  • Executable scriptline 1
    Executable file

skills/evolve/scripts/evolve_core/sampling_config.py

  • Executable scriptline 1
    Executable file

skills/evolve/scripts/evolve_core/structures.py

  • Executable scriptline 1
    Executable file

skills/evolve/scripts/evolve_core/vector_index.py

  • Executable scriptline 1
    Executable file

How the signals fit together

Facts from a static read of the files at the pinned commit. They describe what the code does, not why.

Some files could not be read; the facts below still hold.

No combined signals at this commit.

Works with

Agent compatibility, from the files in this repository
AgentStatusEvidenceInstall to
Claude CodeCompatibleStandard SKILL.md skills/evolve/SKILL.md.claude/skills/<skill-name>/
CodexCompatibleStandard SKILL.md skills/evolve/SKILL.md.agents/skills/<skill-name>/
CursorCompatibleStandard SKILL.md skills/evolve/SKILL.md.cursor/skills/<skill-name>/
Gemini CLICompatibleStandard SKILL.md skills/evolve/SKILL.md.gemini/skills/<skill-name>/
GitHub CopilotCompatibleStandard SKILL.md skills/evolve/SKILL.md.github/skills/<skill-name>/
OpenCodeCompatibleStandard SKILL.md skills/evolve/SKILL.md.opencode/skills/<skill-name>/
  • Native The repository ships a manifest, folder or frontmatter field for this agent.
  • Compatible A standard SKILL.md that this agent can read.
  • Unknown No evidence either way.

Derived automatically from paths, manifests and frontmatter. Not tested on each agent.

You might also use

mattpocock/skills

Skills for Real Engineers. Straight from my .agents directory.

316.8kGitHub stars

obra/superpowers

An agentic skills framework & software development methodology that works.

313.7kGitHub stars

NousResearch/hermes-agent

The agent that grows with you

261.2kGitHub stars

multica-ai/andrej-karpathy-skills

Behavioral guidelines to reduce common LLM coding mistakes. Use when writing, reviewing, or refactoring code to avoid overcomplication, make surgical changes, surface assumptions,…

230.3kGitHub stars

anthropics/skills

Public repository for Agent Skills

188.2kGitHub stars

DietrichGebert/ponytail

Makes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote.

185.9kGitHub stars