grant-review

Name: grant-review
Author: neuromechanist

37stars7forks

How to install grant-review?

neuromechanist/research-skills/grant-review

$npx -y skills add neuromechanist/research-skills --skill grant-review

Installs into the current project.

Summary

This skill should be used when the user asks to "review a grant", "review my proposal", "score this grant", "evaluate my specific aims", "critique my research strategy", "review as an NIH reviewer", "review as an NSF panelist", "give me reviewer feedback", "check my grant proposa

Files · 1

View on GitHub

SKILL.md

1	# Grant Proposal Review
2
3	Routes a grant proposal to an independent, fresh-context reviewer that scores it against the official NIH or NSF criteria and returns a structured review. This skill is a thin dispatcher: it decides how to run the reviewer and in which mode. The review procedure, criteria, scoring rubrics, and output templates all live in `references/` and are loaded by the reviewer, not duplicated here.
4
5	## When to use
6
7	Activate when the user wants feedback on a grant proposal (specific aims, research strategy, project description) evaluated against NIH or NSF review criteria.
8
9	## Why a fresh-context reviewer
10
11	Review validity depends on independence: a reviewer that shares the conversation that produced the proposal is biased toward it. Run the reviewer in a separate context and pass only framing (proposal path, mechanism/agency, resubmission status, target program), never the authoring rationale. This is also why the reviewer is a subagent on tools that support one, and an inline procedure where they do not.
12
13	## Modes (user decides each run)
14
15	- Single (default): one independent reviewer applies the full procedure end to end.
16	- Panel (opt-in): spawn N independent reviewers in parallel, then a synthesis pass. Trigger on "mock study section", "panel review", or an explicit request for multiple reviewers. NIH: 3 reviewers; NSF: 2-3 panelists. Each reviewer scores independently from `references/`; a final chair pass reconciles them into one output. Surface genuine disagreement rather than blindly averaging scores.
17
18	## Dispatch
19
20	Pick the branch for the current tool. In every branch the reviewer follows `references/review-procedure.md`.
21
22	- Claude Code: `Task(subagent_type: "grant-review", ...)` passing the proposal path, mechanism, and mode. For panel mode, launch one `Task` per reviewer role in parallel, then a final synthesis `Task`.
23	- Codex CLI: plugin installation exposes this skill, not a Codex subagent. To use a fresh-context Codex reviewer, first copy `agents/templates/grant-review.toml` to `~/.codex/agents/` or `.codex/agents/`, then invoke that configured agent if the current Codex surface supports `/agent`. For panel mode, ensure `max_threads` covers the reviewer count. If no Codex subagent is configured or available, use the fallback branch.
24	- Copilot CLI: plugin installation exposes this skill and, through `.github/plugin/plugin.json`, the `.agent.md` reviewer in `agents/templates/`. Invoke that configured agent when the current Copilot surface supports custom agents; use `/fleet` for panel mode when available. If running outside a plugin install, copy `agents/templates/grant-review.agent.md` to `.github/agents/` or `~/.copilot/agents/`. If no custom agent is available, use the fallback branch.
25	- Fallback (no subagent support, or the user wants an interactive in-thread review): first locate the rubric (`$CLAUDE_PLUGIN_ROOT/skills/grant-review/references`, else `find . -type d -path '*/skills/grant-review/references' \| head -1`); if it cannot be found, stop and tell the user to install the grant plugin rather than reviewing from memory. Then follow `references/review-procedure.md` directly in this context.
26
27	## The brain (do not duplicate into dispatch or agent shells)
28
29	- `references/review-procedure.md` -- step-by-step procedure: mechanism ID, ingest, score, synthesize, output.
30	- `references/nih-review-criteria.md`, `references/nih-career-training-criteria.md`, `references/nsf-review-criteria.md` -- criteria and scoring.
31	- `references/review-best-practices.md` -- calibration and common reviewer comments.
32	- `references/review-output-templates.md` -- NIH and NSF output format.
33	- `examples/sample-nih-r01-review.md` -- worked review; `examples/sample-r01-aims.md` -- sample proposal input for testing.
34	- Sister skill `manuscript:humanizer` -- AI-writing patterns to flag in grant prose.

People who liked this also viewed

6 picks

Type

grant-review

37stars7forks

View on GitHub

How to install grant-review?

neuromechanist/research-skills/grant-review

$npx -y skills add neuromechanist/research-skills --skill grant-review

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# Grant Proposal Review
2
3	Routes a grant proposal to an independent, fresh-context reviewer that scores it against the official NIH or NSF criteria and returns a structured review. This skill is a thin dispatcher: it decides how to run the reviewer and in which mode. The review procedure, criteria, scoring rubrics, and output templates all live in `references/` and are loaded by the reviewer, not duplicated here.
4
5	## When to use
6
7	Activate when the user wants feedback on a grant proposal (specific aims, research strategy, project description) evaluated against NIH or NSF review criteria.
8
9	## Why a fresh-context reviewer
10
11	Review validity depends on independence: a reviewer that shares the conversation that produced the proposal is biased toward it. Run the reviewer in a separate context and pass only framing (proposal path, mechanism/agency, resubmission status, target program), never the authoring rationale. This is also why the reviewer is a subagent on tools that support one, and an inline procedure where they do not.
12
13	## Modes (user decides each run)
14
15	- Single (default): one independent reviewer applies the full procedure end to end.
16	- Panel (opt-in): spawn N independent reviewers in parallel, then a synthesis pass. Trigger on "mock study section", "panel review", or an explicit request for multiple reviewers. NIH: 3 reviewers; NSF: 2-3 panelists. Each reviewer scores independently from `references/`; a final chair pass reconciles them into one output. Surface genuine disagreement rather than blindly averaging scores.
17
18	## Dispatch
19
20	Pick the branch for the current tool. In every branch the reviewer follows `references/review-procedure.md`.
21
22	- Claude Code: `Task(subagent_type: "grant-review", ...)` passing the proposal path, mechanism, and mode. For panel mode, launch one `Task` per reviewer role in parallel, then a final synthesis `Task`.
23	- Codex CLI: plugin installation exposes this skill, not a Codex subagent. To use a fresh-context Codex reviewer, first copy `agents/templates/grant-review.toml` to `~/.codex/agents/` or `.codex/agents/`, then invoke that configured agent if the current Codex surface supports `/agent`. For panel mode, ensure `max_threads` covers the reviewer count. If no Codex subagent is configured or available, use the fallback branch.
24	- Copilot CLI: plugin installation exposes this skill and, through `.github/plugin/plugin.json`, the `.agent.md` reviewer in `agents/templates/`. Invoke that configured agent when the current Copilot surface supports custom agents; use `/fleet` for panel mode when available. If running outside a plugin install, copy `agents/templates/grant-review.agent.md` to `.github/agents/` or `~/.copilot/agents/`. If no custom agent is available, use the fallback branch.
25	- Fallback (no subagent support, or the user wants an interactive in-thread review): first locate the rubric (`$CLAUDE_PLUGIN_ROOT/skills/grant-review/references`, else `find . -type d -path '*/skills/grant-review/references' \| head -1`); if it cannot be found, stop and tell the user to install the grant plugin rather than reviewing from memory. Then follow `references/review-procedure.md` directly in this context.
26
27	## The brain (do not duplicate into dispatch or agent shells)
28
29	- `references/review-procedure.md` -- step-by-step procedure: mechanism ID, ingest, score, synthesize, output.
30	- `references/nih-review-criteria.md`, `references/nih-career-training-criteria.md`, `references/nsf-review-criteria.md` -- criteria and scoring.
31	- `references/review-best-practices.md` -- calibration and common reviewer comments.
32	- `references/review-output-templates.md` -- NIH and NSF output format.
33	- `examples/sample-nih-r01-review.md` -- worked review; `examples/sample-r01-aims.md` -- sample proposal input for testing.
34	- Sister skill `manuscript:humanizer` -- AI-writing patterns to flag in grant prose.

People who liked this also viewed

6 picks

Type