← All skill packs

Research & Analysis Workflows

Skill pack
3 skills0 selected
antondelafuenteSkills in this pack

Plan LLM fine-tuning and evaluation experiments. Use when the user wants to design a new experiment, plan training runs, or create an experimentsummary.yaml file.

English

Ship a scaffold/product change through the SWE pipeline as a GitHub-backed lifecycle: Issue → worktree branch → namespaced design doc → draft PR → cross-family --scaffold design review (posted to the PR) → implement → cross-family --code review (posted) → classifier (records mechanical vs architectural with evidence) → tracked .aar-ci checks + fake-HOME behavior smoke → fail-closed merge-when-clean. Use for any change to the product scaffold (skills, plugins, CI, the constitution). The agents ARE the engineers: a change is authored by one family and reviewed by the OTHER. ENFORCED repos can require the --code review as a native opposite-family engineer review before merge. Worktree-from-the-start — never disturbs the shared main checkout.

English

Use when the user wants to verify marketing claims against evidence — cross-check statistics, awards, certifications, and performance claims with source data.

English

Includes Design Experiment, Ship Change, and Verify Claims.

Install with
Claude CodeCodexCursorGitHub CopilotGemini CLI

Supported by the Atlas install prompt. Skill-specific requirements vary.

Pack stats

Skills3CategoryscienceStars0

What’s included

Design Experiment

Plan LLM fine-tuning and evaluation experiments. Use when the user wants to design a new experiment, plan training runs, or create an experimentsummary.yaml file.

Ship Change

Ship a scaffold/product change through the SWE pipeline as a GitHub-backed lifecycle: Issue → worktree branch → namespaced design doc → draft PR → cross-family --scaffold design review (posted to the PR) → implement → cross-family --code review (posted) → classifier (records mechanical vs architectural with evidence) → tracked .aar-ci checks + fake-HOME behavior smoke → fail-closed merge-when-clean. Use for any change to the product scaffold (skills, plugins, CI, the constitution). The agents ARE the engineers: a change is authored by one family and reviewed by the OTHER. ENFORCED repos can require the --code review as a native opposite-family engineer review before merge. Worktree-from-the-start — never disturbs the shared main checkout.

Verify Claims

Use when the user wants to verify marketing claims against evidence — cross-check statistics, awards, certifications, and performance claims with source data.

Quick start

  1. Choose your skills. Select individual skills above, or choose the entire pack.
  2. Copy the install prompt. Use Copy install prompt above the list.
  3. Send it to your agent. Paste into your coding agent’s chat. Your agent reviews the source and guides the setup.

Try it with your agent

After installation, start with a request like:

Example request
Use the Design Experiment skill to help with my task.
Ask me for the context you need before getting started.

Requirements & license

  • A coding agent that can access your project files and install skills.
  • Your agent checks each skill’s instructions, dependencies, and permissions during setup.
  • Reported license for the skills shown: MIT. The author’s terms apply.
Research & Analysis Workflows | Agent Skills Atlas