/donor-proposal-scoring — score a proposal against the donor grid

One job: take a proposal that has already been written, and say what a donor evaluator would score it and why. Writing proposals is /proposal. Evaluating supplier offers is procurement-offer-review. This is the applicant side, looking at its own submission through the donor eyes.

Mode routing

• SCORE (default) — a call document and a proposal exist; produce a criterion-by-criterion assessment with scores, evidence, a range, and a ranked fix list.
• CALIBRATE — the donor Evaluation Summary Report has arrived and a prior scoring exists; compare them and learn from the gap.

If both documents are present and Ane has not said which mode, ask in one line. Never run CALIBRATE material into a SCORE output: reading the donor first destroys the test.

Two things to settle before anything else

Is this a blind test or live advisory work? They need opposite reading orders.

• Blind test (an ESR exists and the point is to check the method): do not open it. Do not open it to "just check the format". Write the scoring to disk, then compare.
• Live advisory (the proposal is being fixed before submission or resubmission): read every prior ESR from the same donor first. A donor evaluation house style is learnable and it is the cheapest accuracy available. What did this donor praise, what did it always ask for, which populations did it name unprompted, what did it never mention. Then score.

Whose proposal is it? If the applicant is IPPF or Ane own team, this is partly self-assessment and independence is limited by construction. Say so in the output, in the opening notes, not in a footnote. Search the proposal for Ane name and say whether she appears in it and in what role. Score from the call wording, not from what you know the organisation meant.