How to Use AI Skills
Skill File
Saved instructions for one M&E job, such as a logframe review.
Your AI App
Claude, ChatGPT or another AI. You add the skill here.
The Task
The AI follows the skill on your files. You check the result.
Claude
Claude Code
- Add the marketplace, then install the review and workspace plugins:
/plugin marketplace add mestudiohq/me-review /plugin install me-review@me-review /plugin install me-workspace@me-review - If the install summary asks you to reload plugins, follow that instruction.
- Type
/me-workspace:or/me-review:to see the skills.
Claude in Your Browser
- Open a task page. When a verified skill ZIP is available, it is there.
- Check the provider’s current instructions before choosing an account or uploading files.
ChatGPT
- Choose an M&E task kit and check its minimum inputs and decisions.
- Select ChatGPT on the task page. Choose the working mode, audience and requested output, then copy the task starter.
- Paste it into a chat. Share only material you are authorized to share. The starter asks for missing inputs before it drafts.
- At each stage, copy that stage’s instructions and check the result before moving on.
- Keep each artifact in your own files. To resume in a new chat, copy the resume instructions and bring the listed artifacts again.
This site does not run ChatGPT or receive your documents. To start now, make a custom GPT with the kit.
Any Other AI
A skill is plain text, so it needs no installed plugin. Copy one complete stage from a task page into any AI you can paste text or attach readable files to.
Share only readable, de-identified material that stage needs, then check the output against your sources before you use it.
Read the Getting Started GuidePlan an Evaluation
- Evaluation QuestionsPrioritize answerable questions tied to the decisions an evaluation must support.
- Question and Method AlignmentMatch evaluation questions to feasible evidence, methods and limitations.
- Evaluation TOR DraftDraft a source-linked TOR with feasible scope and unresolved decisions visible.
- Sampling Plan DraftDraft a sampling plan with visible calculations, or check the sizing and selection of an existing one.
- Baseline PlanningPlan baseline measurement for every indicator, or check a baseline against the framework, without inventing starting values.
- Evaluation Planning(Workflow)Move from evaluation decisions to questions, methods and a revised TOR.
Design Measurement
- Indicator Reference SheetsDefine each indicator so people can collect and interpret it consistently.
- Measurement Feasibility and BurdenCalculate collection workload and compare practical simplification options.
- M&E Plan Draft and CoherenceBuild a practical plan and check that definitions, responsibilities and use agree.
- Results Chain AlignmentCompare a theory of change, logframe and indicators for gaps and conflicts, keeping source versions visible.
- Data Quality Assessment PlanPlan a risk-based data quality assessment with proposed checks, owners and timing.
- Feasible Measurement(Workflow)Move from indicator review to reference sheets and a feasible M&E plan.
Check Documents
- Logframe ReviewCheck results logic, indicators, assumptions and means of verification.
- Theory of Change ReviewExamine the results chain, causal assumptions and evidence behind it.
- Indicator Quality ReviewAssess definitions, SMART and CREAM qualities, disaggregation and measurability.
- M&E Plan ReviewIdentify gaps and inconsistencies in a measurement and learning plan.
- Proposal M&E ReviewReview the approach, evaluation arrangements, resources and staffing in a proposal.
- Evaluation TOR ReviewReview commissioning scope, methodological feasibility, ethics and deliverables.
- Survey ReviewCheck questionnaire quality, structure and data collection feasibility.
- Sampling Plan ReviewReview the sampling approach, assumptions, frame and calculations.
- Baseline Report ReviewReview baseline evidence, indicator coverage and the basis for later comparison.
- Progress Report ReviewCheck reported numbers, explanations and evidence-linked conclusions.
- Data Quality ReviewReview data quality evidence, verification and proposed improvement actions.
- Evaluation Report ReviewReview design, evidence, findings, conclusions and recommendations.
Turn Evidence into Action
- Progress Evidence and ComparisonsReconcile reported results and identify which comparisons the evidence supports.
- Findings to ActionsTurn supported findings into a traceable proposed management response.
- Reporting to Action(Workflow)Move from checked results to a revised report and proposed actions.
Complete an M&E Task
- Build and Check a Results Structure
- Make Measurement Feasible
- Plan a Baseline and Defensible Sample
- Prepare an Instrument for a Real Pilot
- Analyze Survey Evidence Responsibly
- Commission an Answerable Evaluation
- Plan or Review Operational Data Quality
- Design or Review a Quasi-experimental Evaluation
- Assess Value for Money or Cost Effectiveness
- Turn Reporting Evidence into Action
- Connect Learning to Decisions and Use
- Connect Qualitative Evidence to Learning
From the Guide
Read the Full GuideStart One Job
Choose a task, copy one complete stage into Claude or ChatGPT, and supply only the readable, de-identified material that stage needs.
Check the Output
Compare key facts, figures, dates and commitments to your sources. A person who knows the program reads the whole file. Proposed owners, dates and actions are not approvals.
Resume From Saved Work
Keep the source manifest, v1, the review issues and your decisions. A summary or prior chat memory does not replace them.