Guide · Agent skill
How to use the Meta-Analysis Methods agent skill
Updated · written with AI from the project's own documentation and checked against our catalogue data
The Meta-Analysis Methods skill (skill name: meta-analysis) is a short set of written guidelines that tells an AI agent to follow standard meta-analysis practice when it combines results from several studies or experiments. It is written for researchers, students and anyone who wants cross-experiment results presented more rigorously. After this guide you will know what the skill changes in the agent's behaviour, how to install it, how to ask for it in plain language, and where its limits are.
Meta-Analysis Methodsaiming-lab/AutoResearchClaw · MIT Download the skill Quick start
- 1Download the ZIP. Use the download link on this page to get the meta-analysis skill as a ZIP file.
- 2Unzip the folder. Extract the archive so that SKILL.md is inside the meta-analysis folder.
- 3Place the folder. Move it into ~/.claude/skills/ for personal use or into .claude/skills/ in your project.
- 4Or upload it in the Claude apps. In the Claude apps, go to Settings → Capabilities → Skills and upload the ZIP.
- 5Ask in plain language. Request a cross-study comparison and mention effect sizes, pooled results or across studies.
- 6Check the output. Review the effect sizes, intervals and any outlier flags before relying on them.
What the meta-analysis skill makes the agent do
An agent skill is a folder with a SKILL.md file that gives an agent extra instructions for a specific kind of task. This one applies when the agent is working with results spread across several studies or runs. Its description says it is for aggregating cross-study or cross-experiment results.
The skill's guidance asks the agent to follow seven practices when comparing results across studies or experiments.
- Report effect sizes, not just p-values
- Use standardized metrics for cross-study comparison
- Account for heterogeneity (different setups, datasets, seeds)
- Report confidence intervals alongside point estimates
- Use forest plots to visualize cross-study comparisons
- Identify and discuss outliers or inconsistent results
- Consider publication bias when interpreting aggregate results
Install the meta-analysis skill
Download the skill as a ZIP file from the download link on this page, then unzip it. You can install it in one of two ways.
For a personal install, unzip the folder into ~/.claude/skills/. For a single project, unzip it into the project's .claude/skills/ folder. You can also upload the ZIP in the Claude apps under Settings → Capabilities → Skills.
- Personal: ~/.claude/skills/
- Project: .claude/skills/ inside your project
- Claude apps: Settings → Capabilities → Skills, then upload the ZIP
Check that the skill is in place
After unzipping, the skill folder should contain SKILL.md. That is the only file in the package. If the file is missing, the install did not work and the agent will not have the guidance.
The skill's own metadata lists the trigger keywords it responds to: meta-analysis, effect size, pooled, cross-study and aggregat.
- The package contains SKILL.md only
- Trigger keywords include meta-analysis, effect size, pooled, cross-study and aggregat
How to trigger the skill with natural requests
You do not need special syntax. Ask for cross-study work in plain language and include words like effect size, pooled, or across studies. The agent can match the request to the skill's description and keywords.
Good requests name the set of results you want combined and what you want reported.
- Combine the effect sizes from these five experiments and show confidence intervals.
- Aggregate the accuracy results from my three papers and flag any outliers.
- Compare these runs across different random seeds and discuss heterogeneity.
- Make a forest plot of the cross-study results in this CSV and consider publication bias.
What the skill needs and its safety notes
The skill lists no required tools or dependencies. It does not run scripts and does not use the network. The package has no code, only written guidance, so it does not compute statistics for the agent.
Because the agent's math is not verified by the skill, a person should check any calculations, especially before the results are used for medical or policy decisions.
- Requires: nothing listed
- Runs scripts: no
- Uses network: no
- The file contains only written guidance, not a calculator
Tips for getting better results
Give the agent the raw numbers or a clear table, along with how each study was set up. The skill asks the agent to account for differences in setups, datasets and seeds, and it can only do that if it knows them.
Treat the output as a structured draft. Check the effect size formulas and the interval calculations yourself.
- Provide sample sizes, metrics and setup details for each study
- State which seeds or datasets differ between runs
- Ask explicitly for confidence intervals and a forest plot
- Ask the agent to name any outliers and discuss publication bias
What people use it for
- Comparing accuracy or loss results from several experiments run with different seeds
- Summarising effect sizes from multiple papers for a literature comparison
- Preparing a forest plot of cross-experiment results for a report
- Teaching students the reporting habits of cross-study aggregation
Frequently asked questions
How do I install the meta-analysis agent skill?
Download the ZIP from this page and unzip it. Put the folder in ~/.claude/skills/ for personal use or in a project's .claude/skills/ folder. You can also upload the ZIP in the Claude apps under Settings → Capabilities → Skills.
Is the meta-analysis skill free?
The skill is released under the MIT licence, as listed on this page. Check the repository's licence file for the full terms.
What does the meta-analysis skill require?
The skill lists no required dependencies. It does not run scripts and does not use the network. It contains only a SKILL.md file with written guidance.
Does the skill calculate statistics for me?
No. The skill is a short set of guidelines, not a calculator. It contains no code, so any calculations the agent makes must be checked by a person.
Why is my agent not using the meta-analysis skill?
Confirm that SKILL.md sits inside the skill folder in ~/.claude/skills/ or .claude/skills/. Then make your request use trigger words such as meta-analysis, effect size, pooled, cross-study or aggregate.
Can I use this skill for a single experiment?
It is not meant for that. The cross-study steps do not apply to simple single-experiment analysis.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
Use this skill if you compare results from several experiments or papers and want an agent to report effect sizes, intervals, heterogeneity and outliers more rigorously. Pick something else if you need verified statistical computation, a single-experiment analysis, or a basis for medical or policy decisions, since this is written guidance only.
Download the skill