Skip to content
Agentmart
Submit a tool

self-play

Run a clean-room self-play search over a problem space you have not solved yet — an AlphaGo-shaped loop over text where isolated sub-agent players (a blind…

Skill by lilCnid

IndexedNOASSERTION★ 4
Write a review

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/lilcnid-self-play

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

Run a clean-room self-play search over a problem space you have not solved yet — an AlphaGo-shaped loop over text where isolated sub-agent players (a blind gatherer, an adversary, a blind evaluator, a blind judge) probe a candidate design and each iteration picks the next variable to vary. Isolation is what makes the output evidence: each player is blind to the one thing that would let it serve the builder's wish, so your reasoning, intent and stake cannot leak into the evidence or the scoring. Use when "does this actually work / discriminate / hold up?" is open and your own read is not trustworthy evidence — a rubric, admission gate, classifier, router, retrieval strategy, prompt template, judge, summarizer, or agent policy. Triggers include "explore this design space", "red-team this design", "does this rubric hold", "clean-room eval", "validate this without fooling myself." Do NOT use for a deterministic feature directly checkable without a clean room, or when the ceremony costs more than the answer.

Use cases: design, agent-building

Reviews

Write a review

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/self-play

What's inside

Skill name
self-play
Description
Run a clean-room self-play search over a problem space you have not solved yet — an AlphaGo-shaped loop over text where isolated sub-agent players (a blind gatherer, an adversary, a blind evaluator, a blind judge) probe a candidate design and each iteration picks the next variable to vary. Isolation is what makes the output evidence: each player is blind to the one thing that would let it serve the builder's wish, so your reasoning, intent and stake cannot leak into the evidence or the scoring. Use when "does this actually work / discriminate / hold up?" is open and your own read is not trustworthy evidence — a rubric, admission gate, classifier, router, retrieval strategy, prompt template, judge, summarizer, or agent policy. Triggers include "explore this design space", "red-team this design", "does this rubric hold", "clean-room eval", "validate this without fooling myself." Do NOT use for a deterministic feature directly checkable without a clean room, or when the ceremony costs more than the answer.
License
NOASSERTION
Instructions
321 lines in SKILL.md · view on GitHub

Files (11)

  • LICENSE.md
  • SKILL.md
  • COMPLEXITY_CONVOCATION_VALIDATION.md · references/
  • DOUBTS_WORKSPACE.md · references/
  • PRIMITIVE_ENCODING_AUDIT.md · references/
  • README.md · references/
  • discipline-cases.md · references/
  • earth_figure_factbase.md · references/
  • ground-block-failure.md · references/
  • measurement-and-reporting.md · references/
  • measurement-bounds.md · references/

The license isn't open source, so the instructions aren't shown here. Read SKILL.md on GitHub

More from lilCnid/dovetail

Signals

Stars
4
Forks
0
Open issues
0
Last commit (this folder)
Sep 4, 2026 (1 month ago)
Last release
Aug 6, 2026 (2 months ago)

From GitHub (fetched Oct 24, 2026).

Details

License
NOASSERTION
Repository
github.com/lilCnid/dovetail
First seen
Oct 24, 2026
Last checked
Oct 24, 2026
Source
GitHubgithub-graphql

More from lilCnid

Cite this page

self-play skill. Agent Mart. https://agentmart.dev/skills/lilcnid-self-play (last updated 2026-10-24)