Proposal Review
Produce a rigorous, decision-ready review for AI/ML, computational biology, and bioscience proposals. Be fair, skeptical, specific, and explicit about missing information.
Instructions
- Read the proposal and identify the decision context if provided: sponsor goals, rubric, budget cap, timeline, and risk tolerance.
- If critical information is missing, do not invent it. Flag the gap and turn it into a prioritized question for the PI.
- Structure the review with these sections:
- Executive summary
- Heilmeier catechism
- Technical merit
- Data, compute, and experimental resources
- Risk register
- Team and execution capability
- Ethics, safety, and compliance
- Budget and schedule realism
- Scorecard
- Decision and funding conditions
- Questions for the PI
- Tailor the technical review to the proposal type:
- AI/ML: baselines, ablations, leakage prevention, calibration, external validation, compute realism
- Bio or wet lab: controls, replicates, statistical plan, assay feasibility, translational path
- Include at least six risks covering technical, data or experimental, budget or timeline, and adoption or regulatory concerns when relevant.
- If the sponsor supplies a rubric, use its categories, weights, and decision vocabulary. Otherwise use the default 1-to-5 scorecard below; do not mix sponsor and default weights.
- Default weights: strategic fit and novelty 15%, technical rigor 25%, feasibility and resources 20%, team and execution 15%, risk, ethics, and compliance 15%, budget and schedule 10%.
- Map the default weighted mean to
Strong Accept (>=4.5), Accept (>=3.7), Borderline (>=2.8), or Reject (<2.8). A documented fatal flaw may override the numeric band.
- Keep the review concrete and action-oriented. Reference proposal details when available and name fatal flaws plainly.
- For a machine-checked scorecard, run
uv run --script skills/proposal-review/scripts/score_proposal.py scorecard.json. The
helper rejects weights that do not total 100%, category mismatches, and
scores outside 1–5. Sponsor rubrics must provide both weights and their own
recommendation bands, so defaults are never mixed into a sponsor rubric.
Quick Reference
| Task |
Action |
| Summarize proposal |
Describe aims, novelty, and bottom-line recommendation in <=150 words |
| Test strategic logic |
Answer the Heilmeier catechism explicitly |
| Review feasibility |
Check assumptions, methods, milestones, and resource realism |
| Review rigor |
Assess controls, baselines, validation, statistics, and reproducibility |
| Review risk |
Build a risk register with likelihood, impact, warning signs, and mitigations |
| Make a decision |
Give a final recommendation plus concrete funding conditions or rejection reasons |
| Use a sponsor rubric |
Preserve its categories, weights, thresholds, and recommendation labels |
| Validate a scorecard |
uv run --script skills/proposal-review/scripts/score_proposal.py scorecard.json |
Input Requirements
- Proposal text or a linkable proposal excerpt
- Optional sponsor or program context
- Optional scoring rubric, budget cap, and timeline constraints
Output
- A decision-ready structured proposal review
- A weighted scorecard with justified subscores
- A clear funding recommendation and conditions
- A prioritized list of questions that could change the decision
Quality Gates
Examples
Example 1: Review a computational biology grant draft
Review this proposal for a microbiome foundation-model project. Use a 1-5 scorecard,
identify fatal flaws if any, and list conditions for funding.
Example 2: Review with sponsor constraints
Review this translational bioscience proposal for a program with a 24-month timeline,
$1.5M budget cap, and high concern for regulatory risk.
Troubleshooting
Issue: The proposal is missing a clear evaluation plan
Solution: Mark this as a major weakness, explain what convincing evidence would look like, and add PI questions about milestones and success metrics.
Issue: The budget or timeline is hard to judge
Solution: State the uncertainty, identify the likely critical path, and evaluate whether the claimed scope is credible under the stated constraints.
Issue: Ethics or compliance details are absent
Solution: Treat the omission as a potential blocker and ask targeted questions about subjects, privacy, biosafety, or regulatory readiness.
Related Skills
/manuscript-review-council — equivalent pipeline for manuscripts/scientific-writing — draft or revise the proposal narrative/bio-logic — assess methodology and evidence rigor
1---2name: proposal-review3description: Evaluate a proposed AI/ML, computational-biology, or bioscience research project for a funding decision. Use when reviewing a grant, research proposal, or funding application against sponsor criteria, feasibility, budget, and risks.4---56# Proposal Review78Produce a rigorous, decision-ready review for AI/ML, computational biology, and bioscience proposals. Be fair, skeptical, specific, and explicit about missing information.910## Instructions11121. Read the proposal and identify the decision context if provided: sponsor goals, rubric, budget cap, timeline, and risk tolerance.132. If critical information is missing, do not invent it. Flag the gap and turn it into a prioritized question for the PI.143. Structure the review with these sections:15 - Executive summary16 - Heilmeier catechism17 - Technical merit18 - Data, compute, and experimental resources19 - Risk register20 - Team and execution capability21 - Ethics, safety, and compliance22 - Budget and schedule realism23 - Scorecard24 - Decision and funding conditions25 - Questions for the PI264. Tailor the technical review to the proposal type:27 - AI/ML: baselines, ablations, leakage prevention, calibration, external validation, compute realism28 - Bio or wet lab: controls, replicates, statistical plan, assay feasibility, translational path295. Include at least six risks covering technical, data or experimental, budget or timeline, and adoption or regulatory concerns when relevant.306. If the sponsor supplies a rubric, use its categories, weights, and decision vocabulary. Otherwise use the default 1-to-5 scorecard below; do not mix sponsor and default weights.317. Default weights: strategic fit and novelty 15%, technical rigor 25%, feasibility and resources 20%, team and execution 15%, risk, ethics, and compliance 15%, budget and schedule 10%.328. Map the default weighted mean to `Strong Accept` (>=4.5), `Accept` (>=3.7), `Borderline` (>=2.8), or `Reject` (<2.8). A documented fatal flaw may override the numeric band.339. Keep the review concrete and action-oriented. Reference proposal details when available and name fatal flaws plainly.3410. For a machine-checked scorecard, run35 `uv run --script skills/proposal-review/scripts/score_proposal.py scorecard.json`. The36 helper rejects weights that do not total 100%, category mismatches, and37 scores outside 1–5. Sponsor rubrics must provide both weights and their own38 recommendation bands, so defaults are never mixed into a sponsor rubric.3940## Quick Reference4142| Task | Action |43|------|--------|44| Summarize proposal | Describe aims, novelty, and bottom-line recommendation in <=150 words |45| Test strategic logic | Answer the Heilmeier catechism explicitly |46| Review feasibility | Check assumptions, methods, milestones, and resource realism |47| Review rigor | Assess controls, baselines, validation, statistics, and reproducibility |48| Review risk | Build a risk register with likelihood, impact, warning signs, and mitigations |49| Make a decision | Give a final recommendation plus concrete funding conditions or rejection reasons |50| Use a sponsor rubric | Preserve its categories, weights, thresholds, and recommendation labels |51| Validate a scorecard | `uv run --script skills/proposal-review/scripts/score_proposal.py scorecard.json` |5253## Input Requirements5455- Proposal text or a linkable proposal excerpt56- Optional sponsor or program context57- Optional scoring rubric, budget cap, and timeline constraints5859## Output6061- A decision-ready structured proposal review62- A weighted scorecard with justified subscores63- A clear funding recommendation and conditions64- A prioritized list of questions that could change the decision6566## Quality Gates6768- [ ] Missing information is flagged instead of invented69- [ ] The review covers novelty, rigor, feasibility, risks, team, ethics, and budget70- [ ] At least six concrete risks are documented with mitigations71- [ ] The final recommendation is explicit and consistent with the evidence72- [ ] Scorecard weights total 100%, all rubric categories are scored, and the recommendation follows the selected rubric's bands7374## Examples7576### Example 1: Review a computational biology grant draft7778```text79Review this proposal for a microbiome foundation-model project. Use a 1-5 scorecard,80identify fatal flaws if any, and list conditions for funding.81```8283### Example 2: Review with sponsor constraints8485```text86Review this translational bioscience proposal for a program with a 24-month timeline,87$1.5M budget cap, and high concern for regulatory risk.88```8990## Troubleshooting9192**Issue**: The proposal is missing a clear evaluation plan93**Solution**: Mark this as a major weakness, explain what convincing evidence would look like, and add PI questions about milestones and success metrics.9495**Issue**: The budget or timeline is hard to judge96**Solution**: State the uncertainty, identify the likely critical path, and evaluate whether the claimed scope is credible under the stated constraints.9798**Issue**: Ethics or compliance details are absent99**Solution**: Treat the omission as a potential blocker and ask targeted questions about subjects, privacy, biosafety, or regulatory readiness.100101## Related Skills102103- `/manuscript-review-council` — equivalent pipeline for manuscripts104- `/scientific-writing` — draft or revise the proposal narrative105- `/bio-logic` — assess methodology and evidence rigor
fmschulz