Skip to content
Get packs

Meta-Optimize Harness Tuner

Analyzes ARIS usage logs and proposes improvements to its skills, prompts and defaults

Source Download skill
Repository
wanshuiyin/Auto-claude-code-research-in-sleep
Stars
18,482
Licence
MIT
Files
1 file

What it does

This skill makes an agent review a log of past ARIS research-workflow runs and suggest changes to the instructions, default settings and review rules that ARIS uses to drive its own skills. It counts which skills get used, where runs fail or need human correction, and how many review rounds it takes to converge, then ranks proposed fixes. It is designed to propose changes rather than apply them automatically, and it does not change the research papers or code produced by the workflows.

Great for

  • Developers maintaining the ARIS research harness who keep event logs of skill runs
  • Teams who want evidence-based tuning of review thresholds and retry limits
  • Researchers building their own agent workflows who want a worked example of log-driven prompt improvement

Skip it if

  • Ordinary users who just want to write a paper or run experiments, since it needs an accumulated event log first and is about improving the tooling itself
  • Users without a .aris/meta/events.jsonl file, because the skill stops when that log is missing and logging is not created automatically for Codex installs
  • Anyone expecting automatic, unreviewed rewrites of skills, since the proposals still need human review and the skill's own rules forbid deleting safety, review or provenance steps

Try asking

  • “Optimize ARIS skills based on the last month of usage logs”
  • “Analyze my usage records and tell me which skill fails most often”
  • “Improve the reviewer prompts using my accumulated experience”
  • “Show me where human interventions happen most during auto-review loops”
Step-by-step guide: how to use Meta-Optimize Harness TunerInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 1 original files, the MIT licence and an attribution note.

    meta-optimize.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip meta-optimize.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip meta-optimize.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Run it only after several complete ARIS workflow runs have been logged, so the analysis has enough data
  • Review every proposed SKILL.md change yourself before applying it, and keep the bottleneck log as an append-only history
  • Use the argument to scope the analysis to one target skill rather than all skills to keep proposals focused

What's inside

  • SKILL.md

Redistributed unmodified from wanshuiyin/Auto-claude-code-research-in-sleep under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.