Evaluation Planning
Move from evaluation decisions to questions, methods and a revised TOR.
Free, with limits. These skills produce AI-assisted drafts and reviews. A person who knows the program should check every output against the source documents before it is shared. The examples are fictional and written by hand; they are not real client work or test output.
Claude Code is the checked route. One skill was also checked by browser upload on a paid plan. Anthropic's help page says Free plans can use uploaded skills too, once code execution is turned on; we did not test a Free account. No M&E Studio signup or API key is needed; your Claude account and its limits are separate.
What You Will Get
- questions.md
- matrix.md
- tor-draft.md
- review.md
- tor-revised.md
- change-record.md
What to Provide
- One fictional case below, or equivalent readable source documents
- All component skills listed below
- Original inputs, saved artifacts and handover when resuming
Supply readable files or text. Missing evidence stays explicit; a draft or recommendation is not an approved commitment.
Start the Job
Copy the starter after obtaining its component skills below. Use one fictional case at a time. Save every intermediate file and the handover; resuming needs the original input, actual artifacts and handover in a fresh session.
Read the Starter
Evaluation planning starter pack
This pack helps you complete one M&E job with free instructions and fictional attachments. Your Claude account has its own access and usage limits. Checked in Claude Code. Anthropic's help page says Free and paid plans can use uploaded skills once code execution is turned on; we checked browser upload on a paid plan only. No API key or site signup is needed.
Inputs and start
Supply evaluation-planning.md or your equivalent readable files. examples/normal and examples/difficult are separate fictional cases; never combine them. Browser users upload files; Code users give explicit file paths or a scoped folder. If extraction fails, provide a readable text export. This pack promises Markdown output, not Word, PDF or spreadsheet rendering.
Enable these workspace skills: evaluation-questions, question-method-alignment, evaluation-tor-draft. Also enable the review skill(s) tor-review from me-review. In Code load both plugins and use /me-workspace:<skill> and /me-review:<review>. In a browser upload individual skill packages and name the skill. Copy/paste fallback: paste each relevant SKILL.md in sequence with the files or text it needs; browser behavior still needs separate verification.
Starter: Complete evaluation planning with my supplied documents. Use evaluation-questions to write questions.md, question-method-alignment to write matrix.md, and evaluation-tor-draft to write tor-draft.md. Use me-review:tor-review to write a prioritized review.md of that draft against the supplied constraints. Revise into tor-revised.md and record each material finding, change or unresolved decision in change-record.md. Then synchronize questions.md and matrix.md to the final TOR question IDs, wording, core/optional status, priorities, day allocations and schedule. Do not leave conflicting versions for the user to reconcile. Preserve the original TOR draft separately for comparison. Check every final question has an evidence/method row and all interview and fieldwork dates fall within the supplied fieldwork window, if one is given; distinguish later desk analysis from fieldwork. If a day cap is given, check task days add to it. Keep unresolved choices explicit. No conditional source may become confirmed access. Return the final artifacts in editable Markdown. Separate facts, assumptions and proposed choices. Make unresolved decisions explicit. Treat core/optional status, day allocations and intermediate dates as proposals with a stated basis. If no defensible estimate is possible, leave it unresolved rather than inventing effort or commitments.
Save and resume
Number decisions with the stage prefix (EVAL-D1, EVAL-D2) so each ID means one decision across all files. Save intermediate outputs as you go. If interrupted, use handover-template.md to state the input files/source IDs, completed files, decisions/assumptions, remaining steps and exact restart instruction. In a fresh session provide the original input, completed artifacts and handover. Re-validate inherited facts and status claims against the original sources before copying them forward. A handover is a navigation aid, not authority for approval, consent or measurement definitions. Use three evidence states: explicitly confirmed, explicitly contradicted, or unconfirmed. No supporting record means unconfirmed; it cannot justify keeping a negative status claim. If a resumed claim has no supporting record, mark it unknown rather than keeping it. Distinguish approval status of pre-existing program items from recommendations created in this draft (proposed). Do not rely on earlier chat memory or undocumented author corrections.
Expected files: questions.md, matrix.md, tor-draft.md, review.md, tor-revised.md, change-record.md. Check that you can open each file, follow source IDs back to inputs, distinguish proposals from commitments, and see the next decision. Human usability validation remains pending.
Required Skills and Reviews
The workflow pack does not contain these methods. Install the two plugins or download the named individual ZIPs, then follow the starter.
Try a Fictional Example
Choose a complete, difficult or single-period case separately. Never combine alternative cases.
Complete Fictional Case
Evaluation planning case: BridgeWorks
Fictional case. BridgeWorks is an invented youth employment program. All organizations, locations, figures, and records below are fictional and supplied only for this exercise.
Request
Prepare a practical evaluation question and method matrix and a draft terms of reference (TOR) for the program director. The director must decide by December 15, 2026 whether to continue the next cohort as designed or make changes before recruitment. The evaluation should be useful for that decision. The director has authorized up to 20 evaluator working days, including analysis and reporting. Fieldwork can take place during November 2026. No external comparison group is available.
Source BW-01: Program brief
BridgeWorks ran one 2025 cohort at four community centers. It enrolled 160 young adults ages 18 to 24. The program defines completion as attending at least 10 of 12 sessions. The register records 128 people meeting that definition. Sessions covered job search, interview practice, and workplace communication. The program director is the decision owner. The next cohort is planned to begin in January 2027.
Source BW-02: Results framework
| ID | Result or indicator | Definition / available source | 2025 value |
|---|---|---|---|
| BW-O1 | Participants complete the program | Count of enrolled participants attending at least 10 of 12 sessions; attendance register | 128 of 160 enrolled (80%) |
| BW-O2 | Participants use job-search skills | No routine measure collected | Not measured |
| BW-OC1 | Participants obtain paid work within six months | No follow-up system; no comparison group | Not measured |
The program's original planning target for BW-O1 was 75% of enrolled participants. No target was set for the other results.
Source BW-03: Available records
- Attendance register with participant IDs, center, session attendance, and completion status for all 160 enrollees.
- A brief end-of-cohort feedback form completed by 72 participants. It asked whether sessions were useful and what participants would change. The form did not ask about employment status.
- No routine employment follow-up, participant phone list, or consent for recontact is available to the evaluator.
- Four facilitators and the program director can participate in interviews. The director can arrange introductions to up to six local employers, but employer participation is not guaranteed.
Source BW-04: Evaluation constraints
- Evaluator level of effort: maximum 20 working days, inclusive of preparation, fieldwork, analysis, and reporting.
- Fieldwork window: November 1 to November 30, 2026.
- Final report needed by December 10, 2026, so the director can decide by December 15.
- No external comparison group, baseline measure of employment, or routine six-month follow-up exists.
- Findings should distinguish what the available records show from what interviews can suggest.
User task
Using only these supplied sources, produce a prioritized evaluation question and method matrix, a concise draft TOR, and a short review/change record explaining any scope choices. Cite source IDs beside material facts. State assumptions and unresolved limitations. Do not invent evaluation findings or claim that the program caused employment outcomes.
Difficult Fictional Case
Evaluation planning case: Northstar Pathways
Fictional case. Northstar Pathways is an invented rural youth livelihoods program. All organizations, locations, figures, and records below are fictional and supplied only for this exercise.
Request
Draft an evaluation TOR and questions for the steering group. The current request email says the evaluation is for changes to the next cycle, which starts November 16, 2026. The grant summary says the purpose is to decide whether to expand the model to three new districts in 2027. No one is identified as the final decision owner. The request email asks for a final report by November 10, 2026. The grant summary asks for the report by January 31, 2027.
Source NP-01: Program brief
Northstar enrolled 240 participants across three districts in 2025. The program tracks attendance but did not collect a baseline measure of income or employment. The participant register has 240 IDs and district fields. The program manager reports that 174 participants attended at least 8 of 10 sessions. There is no routine follow-up after the final session.
Source NP-02: Draft results framework
| ID | Result or indicator | Definition / source | Stated value |
|---|---|---|---|
| NP-O1 | Participants complete training | Attendance at 8 of 10 sessions; register | 174 of 240 (72.5%) |
| NP-OC1 | Participants increase income by 20% | Follow-up income compared with baseline; no baseline or follow-up tool is listed | Target: 70% |
| NP-OC2 | Participants obtain paid work | Number in paid work six months after training; no follow-up records | Target: 60% |
The framework describes NP-OC1 as a six-month outcome. A note below the table says the 70% target is to be reported at the end of training.
Source NP-03: Steering group notes
- “Show the impact of the model and recommend scale-up.”
- “Use a comparison group if possible; all participants should still receive the full program.”
- “We need an answer before the next cycle begins.”
- No comparison group is currently identified. No random assignment is planned.
- The program manager can share attendance records. The manager has not confirmed whether participant contact details may be shared.
Source NP-04: Resources and schedule
- Available evaluator effort: 12 working days total, including design, fieldwork, analysis, and reporting.
- Available interview window: October 26 to November 6, 2026.
- Request email: final report due November 10, 2026.
- Grant summary: final report due January 31, 2027.
- Steering group meeting: November 12, 2026.
- A separate planning note proposes 120 participant interviews, 30 employer interviews, and a comparison survey in all three districts. No sample frame, survey instrument, or travel plan is attached.
User task
Using only these supplied sources, produce a prioritized evaluation question and method matrix, a draft TOR that can be acted on, and a short review/change record. Cite source IDs beside material facts. Separate what can be designed now from decisions or permissions that must be resolved. Identify incompatible purposes, timing, resources, and claims. Do not make up baseline data, comparison evidence, participant consent, or impact findings. If a decision is needed before a defensible TOR can be finalized, state the smallest decision set and provide a bounded interim option.
Files for This Job
Evidence and Limits
Checked in Claude Code on fictional cases, including incomplete and unsuitable inputs, with the outputs reviewed by a second model. A browser upload check has passed for one skill (Evaluation Questions) on a paid plan. The Anthropic help page says Free plans can use uploaded skills; we did not test a Free account. Practitioner feedback is not checked yet.
Check the actual output against your source documents, especially numbers, missing data, proposed owners and approval status. Treat the result as an AI-assisted draft. Markdown and text tables are the checked output; Word, PDF and spreadsheet output are not promised. Keep each method's rating rules distinct from site rubric scores.
Source revision 5a6819f · Checked 2026-10-26