Skip to content

K-Dense-AI/scientific-agent-skills / skills/research-grants

research-grants

Verified

Supports research proposal preparation and review for NSF, NIH, DOE, DARPA, and Taiwan NSTC, including opportunity-specific requirements, aims, review criteria, budgets, broader impacts, forms, and resubmissions. Use for investigator-authored grant development, compliance matrices, and proposal critiques.

references/review_criteria.md

View on GitHub
# Review criteria by agency

Reviewed 2026-10-01. Extract the actual opportunity's rubric before critiquing a
proposal. Do not average unrelated factors, invent weights or infer fundability from
a generic score. Use a table linking each criterion to the applicant's evidence,
remaining weakness and proposed revision.

## NSF

[Intellectual Merit and Broader Impacts](https://www.nsf.gov/policies/pappg/24-1/ch-3-proposal-processing-review)
both receive full consideration; this is not a fixed 50/50 arithmetic weighting.
Assess knowledge advancement, originality, design, team/resources, societal benefit
and assessment, together with any additional solicitation criteria. Broader impacts
can arise through the research itself. See [nsf_guidelines.md](nsf_guidelines.md).

## NIH

For applicable RPG due dates on/after January 25, 2025, the
[simplified framework](https://www.grants.nih.gov/policy-and-compliance/policy-topics/peer-review/simplifying-review/framework)
uses Importance of the Research (significance/innovation) and Rigor and Feasibility
(approach), each scored 1-9, plus Expertise and Resources assessed for sufficiency.
All inform overall impact. Check the mechanism: not every NIH application uses this
framework. Research Strategy headings remain those required in the application guide.

Additional review criteria can influence impact despite having no separate score;
these differ from considerations such as budget that do not determine scientific merit.
Match protections, inclusion, clinical-trial timeline and other checks to the NOFO.
Current modified review practices can leave competitive applications undiscussed;
see [nih_guidelines.md](nih_guidelines.md). Do not convert a score or percentile to
predicted funding odds or use old paylines as present policy.

## DOE

Use the NOFO's scientific/technical merit, method, personnel/facilities, budget,
mission relevance, programmatic selection factors and any defined weighting/order.
Do not apply an ARPA-E impact rubric to an SC basic-science application. Distinguish
concept-paper assessment from full-application review; see [doe_guidelines.md](doe_guidelines.md).

## DARPA

Use the solicitation's actual evaluation factors and award instrument. Technical
merit, national-security relevance and cost realism commonly matter, but their exact
role is solicitation-specific. Test goals against baselines, credible measurements,
risks, milestones and transition evidence. See [darpa_guidelines.md](darpa_guidelines.md).

## Taiwan NSTC

Use the division/program's published criteria and current application forms.
Innovation, feasibility, investigator capability and contribution are useful drafting
questions, not a universal numeric rubric or passing-score threshold. Differentiate
project outputs from claimed societal outcomes; see [nstc_guidelines.md](nstc_guidelines.md).
add-product-e2e-evalpaperclipai/paperclipAdd or extend a Paperclip full-stack runner E2E workflow, fixture, matcher, or report evidence path for local or Daytona execution.Stars: 115kUpdated Claude CodeCodexVerifiedInstall
add-runner-evalpaperclipai/paperclipAdd or extend a Paperclip Runner protocol evaluation definition, roster, assertion, or report fixture with provenance and narrow validation.Stars: 115kUpdated Claude CodeCodexVerifiedInstall
diagnose-why-work-stoppedpaperclipai/paperclipCompatibility entry point for failed, stalled, looping, or over-recovered Paperclip work. Use fix-failed-paperclip-run for the full diagnosis, regression, repair, real verification, and green-PR workflow; honor an explicitly requested diagnosis-only scope.Stars: 115kUpdated Claude CodeCodexVerifiedInstall
fix-failed-paperclip-runpaperclipai/paperclipDiagnose and fix failed, stuck, slow, or incorrectly blocked Paperclip tasks and runs. Use with a task/run URL or failure report to trace the root cause, prove a regression red then green, check second-order effects, verify the real workflow, and deliver a PR with passing checks.Stars: 115kUpdated Claude CodeCodexVerifiedInstall