grants

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

Grants — NIH Funding Intelligence

基金申请 — NIH资助情报系统

Portability: Requires
bash_tool
(for RePORTER POST via curl), Node.js with
docx
package, and a Consensus MCP connection. Works in Claude Code CLI natively. In Claude.ai with Code Execution + Consensus MCP, the workflow is supported but slower.

Scope: NIH-only. Non-NIH funders (PCORI, DOD CDMRP, VA, foundations) are out of scope and flagged at intake.

For a clinical researcher with a research idea, produce a strategic NIH funding overview as an editable

.docx

. Output covers research positioning analysis, institute mapping, targeted grant discovery, and strategic recommendations the researcher can edit, copy from, and share with their mentor.

可移植性： 需要
bash_tool
（用于通过curl发送RePORTER POST请求）、安装了
docx
包的Node.js，以及Consensus MCP连接。在Claude Code CLI中可原生运行。在开启代码执行+Consensus MCP的Claude.ai中支持该工作流，但速度较慢。

范围：仅NIH。 非NIH资助方（PCORI、DOD CDMRP、VA、基金会）不在服务范围内，会在信息采集环节标记。

针对有研究想法的临床研究者，生成可编辑的

.docx

格式NIH资助策略概述。输出内容涵盖研究定位分析、研究所匹配、目标基金挖掘，以及研究者可编辑、复制并与导师共享的策略建议。

Agent Integrity Rules (Research-Pack Convention)

Agent完整性规则（Research-Pack约定）

Inherited; locked verbatim per PR #657 audit.

Execution discipline. A step isn't complete until result is confirmed received. Consensus calls sequential with 1+ sec pause. RePORTER calls sequential.
Data sourcing. Count only what tool calls returned this session. Never supplement with training knowledge. Training knowledge labeled
```
[Not from Consensus/RePORTER — reference information]
```
and excluded from counts.
Counts & attribution. Queries sent / results shown / results cited — three separate numbers, never conflate. Every cited paper has retrievable URL from this session.
Error handling. On failure → wait 3s → retry once → log. After 3 consecutive failures across tools: stop, alert researcher, explain what's missing. Never silently skip.
Transparency. Audit Log section in the DOCX. Same standards in chat summary as in document.

See

references/reporter_post_patterns.md

for the RePORTER POST canon + plan-tier detection.

继承自PR #657审核的固定规则，不得修改。

执行纪律： 步骤完成的标志是确认收到结果。Consensus调用需按顺序执行，且每次调用间隔1秒以上。RePORTER调用同样按顺序执行。
数据源： 仅统计本次会话中工具调用返回的数据。绝不使用训练数据补充。训练数据需标记为
```
[Not from Consensus/RePORTER — reference information]
```
并排除在统计之外。
统计与归因： 发送的查询数/展示的结果数/引用的结果数——三个独立数字，绝不混淆。每篇引用的论文需包含本次会话可获取的URL。
错误处理： 失败后→等待3秒→重试一次→记录日志。连续3次工具调用失败：停止操作，提醒研究者，说明缺失内容。绝不静默跳过。
透明度： DOCX中包含审计日志部分。聊天摘要与文档遵循相同标准。

查看

references/reporter_post_patterns.md

获取RePORTER POST标准格式及计划层级检测方法。

Phase 1: Grill-Me Intake (6 forcing questions, one at a time)

第一阶段：Grill-Me信息采集（6个强制问题，逐一提问）

Q1 (root) — Research idea

Q1（核心）——研究想法

Describe the research idea in 2–3 sentences. What's the question, what's new, and what's the clinical relevance? Vague answers ("AI for healthcare", "biomarkers for disease X") will be rejected — push for specificity.

Why I'm asking: Five Consensus searches (established / stakes / current approaches / adjacent methods / gaps) depend on a precise research idea. Vague ideas produce vague gap quotes and useless positioning narrative.

Refuse mush. Re-ask once with examples if user is too broad.

用2-3句话描述研究想法。包括研究问题、创新点及临床相关性。模糊答案（如“AI用于医疗”、“疾病X的生物标志物”）将被拒绝——请提供具体内容。

提问原因： 五次Consensus搜索（已有的研究/研究意义/当前方法/相关技术/研究空白）依赖精准的研究想法。模糊的想法会产生模糊的空白引用和无用的定位描述。

拒绝模糊表述。若用户回答过于宽泛，可提供示例后重新提问一次。

Q2 (depends on Q1) — Career stage

Q2（基于Q1）——职业阶段

Career stage — pick one:

Pre-doctoral (PhD student, T32 trainee)

Postdoctoral fellow (F32, K99 candidate)

Early career (K-award candidate, first R01)

Independent investigator (multiple R01s, established lab)

Senior PI (R35, P-series, U01 leadership)

Why I'm asking: Career stage filters mechanism recommendations. F-series for trainees, K-series for early career, R-series for independent. Picking the wrong stage produces unfundable mechanism suggestions.

Forcing choice.

职业阶段——选择一项：

博士前（博士生、T32受训人员）

博士后研究员（F32、K99候选人）

早期职业阶段（K类资助候选人、首次申请R01）

独立研究者（多项R01资助、实验室已建立）

资深PI（R35、P系列、U01项目负责人）

提问原因： 职业阶段会过滤机制建议。受训人员适用F系列资助，早期职业阶段适用K系列，独立研究者适用R系列。选错阶段会产生无法获得资助的机制建议。

必须选择一项。

Q3 (depends on Q2) — Preliminary data status

Q3（基于Q2）——初步数据状态

Preliminary data — pick one:

None (de novo project, no pilot data yet)

Pilot data (early findings, single-site)

Strong preliminary (multi-experiment, ready for R01-scale)

Validated and ready (multi-site, publication-ready)

Why I'm asking: Prelim data status drives mechanism budget. No data → R03 / R21 pilot scope. Strong prelim → R01 / U01 multi-site scale. Mismatch produces uncompetitive applications.

初步数据状态——选择一项：

无（全新项目，无试点数据）

试点数据（早期结果，单中心）

充足初步数据（多组实验，达到R01规模要求）

已验证就绪（多中心，可发表）

提问原因： 初步数据状态决定机制预算。无数据→R03/R21试点规模；充足初步数据→R01/U01多中心规模。不匹配会导致申请缺乏竞争力。

Q4 (depends on Q2) — Environment

Q4（基于Q2）——研究环境

Research environment — pick one:

R01-eligible (research-intensive institution with NIH base funding)

Mid-tier (regional academic medical center, modest NIH portfolio)

Resource-constrained (smaller institution, minimal NIH base)

Industry-collaborative (academic + industry partnership)

Why I'm asking: Environment affects scope realism (multi-site U01 requires R01-eligible) and which mechanism categories are competitive (R15 specifically targets resource-constrained).

研究环境——选择一项：

符合R01申请资格（研究密集型机构，有NIH基础资助）

中等层级（区域学术医疗中心，NIH资助组合规模适中）

资源受限（小型机构，NIH基础资助极少）

校企合作（学术+产业合作）

提问原因： 研究环境影响范围的现实性（多中心U01项目需符合R01申请资格），以及哪些机制类别具有竞争力（R15专门针对资源受限机构）。

Q5 (depends on Q1) — Submission posture

Q5（基于Q1）——提交状态

Submission posture — pick one:

New application (first submission, no prior reviews)

Resubmission (A1 with reviewer responses needed)

Exploring (haven't decided yet whether to submit)

Why I'm asking: Resubmissions need reviewer-response guidance in the DOCX (Section 7). New applications skip that. Exploring shifts emphasis to landscape over strategy.

提交状态——选择一项：

新申请（首次提交，无过往评审意见）

重新提交（A1版本，需回应评审意见）

探索阶段（尚未决定是否提交）

提问原因： 重新提交的申请需要在DOCX中包含评审意见回应指导（第7节）。新申请可跳过该部分。探索阶段会将重点从策略转向整体格局。

Q6 (depends on Q1) — Known institute targets

Q6（基于Q1）——目标研究所

Are you already considering specific NIH institutes? List names (NCI / NHLBI / NIMH / NINDS / NIDDK / etc.) or say "no preference — find the right ones".

Why I'm asking: If you have an institute hypothesis, I'll validate it against RePORTER data. If not, I'll surface the top-3 institutes funding adjacent work from the institute-tally.

Accept "no preference" as the common case.

Stop condition: After Q6, commit and start Phase 2A. Never re-open intake after Phase 2A begins.

你是否已考虑特定的NIH研究所？列出名称（NCI/NHLBI/NIMH/NINDS/NIDDK等）或回答“无偏好——帮我找到合适的”。

提问原因： 如果你已有目标研究所的设想，我会通过RePORTER数据验证。如果没有，我会从研究所统计数据中找出资助相关研究最多的前3个研究所。

接受“无偏好”作为常见选项。

停止条件： Q6完成后，确认并启动第二阶段A。第二阶段A开始后不得重新开启信息采集环节。

Phase 2A: Research Positioning (5 Consensus searches)

第二阶段A：研究定位（5次Consensus搜索）

Run sequentially at 1 q/sec. Each search corresponds to one positioning facet:

Established —
```
"<research idea>" established evidence
```
— what's known

Stakes —

"<topic>" mortality OR burden OR cost OR prevalence

— why it matters

Current Approaches —

"<topic>" current treatment OR standard of care OR approach

— state of the art

Adjacent Methods —
```
"<related technique>" applied to <topic>
```
— methodological possibilities

Gaps —

"<topic>" limitations OR unanswered OR future directions OR challenge

— gap signals

Use

scripts/citation_tracker.py --action record_consensus_search

for each. Plan-tier detected from first response.

Synthesis: for each facet, extract 2-3 quotable findings (becomes Section 2 gap quotes). Draft Significance/Innovation language using "the field has established X (refs), but Y remains unanswered (refs)" pattern.

按每秒1次的频率顺序执行。每次搜索对应一个定位维度：

已有的研究 —
```
"<research idea>" established evidence
```
— 已知内容

研究意义 —

"<topic>" mortality OR burden OR cost OR prevalence

— 重要性

当前方法 —

"<topic>" current treatment OR standard of care OR approach

— 现有技术水平

相关技术 —

"<related technique>" applied to <topic>

— 方法学可能性

研究空白 —

"<topic>" limitations OR unanswered OR future directions OR challenge

— 空白信号

每次搜索使用

scripts/citation_tracker.py --action record_consensus_search

。从首次响应中检测计划层级。

合成： 针对每个维度，提取2-3个可引用的发现（成为第2节的空白引用）。使用“领域已证实X（参考文献），但Y仍未解决（参考文献）”的模式撰写重要性/创新性内容草稿。

Phase 2B: Institute Mapping + Grant Discovery (RePORTER POST)

第二阶段B：研究所匹配 + 基金挖掘（RePORTER POST）

RePORTER is POST-only. Use

bash_tool

curl

— never

web_fetch

RePORTER仅支持POST请求。使用

bash_tool

curl

— 绝不使用

web_fetch

。

Dynamic fiscal year window

动态财年窗口

Compute at runtime via

scripts/fiscal_year_calculator.py

. Default: current FY + 3 prior. Federal FY starts Oct 1, so:

bash

python ../scripts/fiscal_year_calculator.py --output json

通过

scripts/fiscal_year_calculator.py

实时计算。默认：当前财年+前3个财年。联邦财年从10月1日开始，示例：

bash

python ../scripts/fiscal_year_calculator.py --output json

Returns: {"current_fy": 2026, "window": [2023, 2024, 2025, 2026]}

返回: {"current_fy": 2026, "window": [2023, 2024, 2025, 2026]}

undefined

undefined

Narrow (AND) search — finds direct overlap

精确（AND）搜索——查找直接重叠项目

bash

curl -X POST 'https://api.reporter.nih.gov/v2/projects/search' \
  -H 'Content-Type: application/json' \
  -d '{
    "criteria": {
      "fiscal_years": [2023, 2024, 2025, 2026],
      "include_active_projects": true,
      "advanced_text_search": {
        "operator": "AND",
        "search_field": "all",
        "search_text": "<key term 1> <key term 2>"
      }
    },
    "limit": 50,
    "include_fields": ["project_num", "project_title", "agency_ic_admin", "study_section", "fiscal_year", "principal_investigators", "abstract_text"]
  }'

bash

curl -X POST 'https://api.reporter.nih.gov/v2/projects/search' \
  -H 'Content-Type: application/json' \
  -d '{
    "criteria": {
      "fiscal_years": [2023, 2024, 2025, 2026],
      "include_active_projects": true,
      "advanced_text_search": {
        "operator": "AND",
        "search_field": "all",
        "search_text": "<key term 1> <key term 2>"
      }
    },
    "limit": 50,
    "include_fields": ["project_num", "project_title", "agency_ic_admin", "study_section", "fiscal_year", "principal_investigators", "abstract_text"]
  }'

Broad (OR) search — finds adjacent work

宽泛（OR）搜索——查找相关项目

bash

curl -X POST 'https://api.reporter.nih.gov/v2/projects/search' \
  -H 'Content-Type: application/json' \
  -d '{
    "criteria": {
      "fiscal_years": [2023, 2024, 2025, 2026],
      "advanced_text_search": {
        "operator": "OR",
        "search_field": "all",
        "search_text": "<term> <synonym> <related concept>"
      }
    },
    "limit": 50
  }'

bash

curl -X POST 'https://api.reporter.nih.gov/v2/projects/search' \
  -H 'Content-Type: application/json' \
  -d '{
    "criteria": {
      "fiscal_years": [2023, 2024, 2025, 2026],
      "advanced_text_search": {
        "operator": "OR",
        "search_field": "all",
        "search_text": "<term> <synonym> <related concept>"
      }
    },
    "limit": 50
  }'

Institute tally + study section ranking

研究所统计 + 评审组排名

After RePORTER responses:

Tally
```
agency_ic_admin
```
(institute code: NCI, NHLBI, NIMH, etc.) → top-3 funding institutes
Tally
```
study_section
```
→ top-2 study sections (where applications go for review)

获取RePORTER响应后：

统计
```
agency_ic_admin
```
（研究所代码：NCI、NHLBI、NIMH等）→ 资助最多的前3个研究所
统计
```
study_section
```
→ 排名前2的评审组（申请提交至这些组进行评审）

NOSI discovery

NOSI挖掘

Parse RePORTER responses for

NOT-*

opportunity numbers. For each:

bash

undefined

从RePORTER响应中解析

NOT-*

机会编号。针对每个编号：

bash

undefined

NOSIs live at predictable URLs:

NOSI的URL格式固定：

https://grants.nih.gov/grants/guide/notice-files/NOT-<INSTITUTE>-<YEAR>-<NUMBER>.html

web_fetch <url>


If fetch fails: log `[NOSI {number} — fetch failed, not included]`, continue.

web_fetch <url>


若获取失败：记录`[NOSI {number} — 获取失败，未包含]`，继续执行。

Mechanism Matching (Scope-Aware)

机制匹配（适配范围）

NOT career stage alone. Career stage + project scope + prelim data drive recommendation.

Use

scripts/mechanism_matcher.py

bash

python ../scripts/mechanism_matcher.py \
  --career-stage "early_career" \
  --prelim-data "pilot" \
  --environment "r01_eligible" \
  --scope "single_site" \
  --output json

并非仅基于职业阶段。职业阶段**+项目范围+**初步数据共同驱动推荐。

使用

scripts/mechanism_matcher.py

：

bash

python ../scripts/mechanism_matcher.py \
  --career-stage "early_career" \
  --prelim-data "pilot" \
  --environment "r01_eligible" \
  --scope "single_site" \
  --output json

Returns mechanism shortlist with rationale

返回包含理由的机制候选列表


See [`references/nih_mechanism_matching.md`](references/nih_mechanism_matching.md) for the full matrix.


查看[`references/nih_mechanism_matching.md`](references/nih_mechanism_matching.md)获取完整匹配矩阵。

Phase 3: DOCX Generation

第三阶段：DOCX生成

9 sections via Node.js +

docx

library. See

references/docx_9_sections.md

for full spec.

Executive Summary — title + career stage + environment + 3-4 key findings bullets
Research Positioning — 3-5 gap quotes (italicized, inline Consensus citations) + 2-3 paragraph positioning narrative + supporting evidence table
Target Institutes — ranking table (institute, project count in window, % match to your idea) + 2-3 sentence interpretation
Grant Opportunities — bold NOSI callout if any. Top-3 grants table with hyperlinked FOAs + per-grant scope/budget fit paragraph
Funded Overlap — top-5 projects table (PI, project_num, IC, year, hyperlinked to RePORTER) + differentiation paragraph
Study Sections — ranking table + best-match interpretation
Strategic Recommendations & Next Steps — 3-4 numbered recs + mandatory program officer rec + submission timeline note + (if resubmission Q5=2) reviewer-response guidance + closing paragraph
References — numbered bibliography, hyperlinked to Consensus
Audit Log — Consensus searches table, plan-tier note, RePORTER searches table, NOSI fetches table, summary stats, tool constraints note, failed steps

通过Node.js +

docx

库生成9个章节。查看

references/docx_9_sections.md

获取完整规范。

执行摘要 — 标题+职业阶段+研究环境+3-4个关键发现要点
研究定位 — 3-5个空白引用（斜体，内嵌Consensus引用）+2-3段定位描述+支持证据表格
目标研究所 — 排名表格（研究所、财年窗口内项目数、与你的想法匹配度百分比）+2-3段解读
基金机会 — 若存在NOSI则突出显示。前3个基金表格，包含带超链接的FOA+每个基金的范围/预算适配说明段落
已资助项目重叠 — 前5个项目表格（PI、项目编号、研究所、年份、RePORTER超链接）+差异化说明段落
评审组 — 排名表格+最佳匹配解读
策略建议与下一步行动 — 3-4条编号建议 + 强制要求的项目官员推荐 + 提交时间表说明 +（若Q5=2为重新提交）评审意见回应指导 + 结尾段落
参考文献 — 编号参考文献目录，带Consensus超链接
审计日志 — Consensus搜索表格、计划层级说明、RePORTER搜索表格、NOSI获取表格、统计摘要、工具限制说明、失败步骤记录

Styling

样式

Arial 12pt body, navy headings (#1a3a5c), light blue table headers (#e8f0f8), amber NOSI callout.

ExternalHyperlink

patterns:

Paper citations:
```
https://consensus.app/papers/...
```
FOA links:
```
https://grants.nih.gov/grants/guide/...
```

RePORTER projects:

https://reporter.nih.gov/project-details/<id>

正文为Arial 12号字体，标题为深蓝色（#1a3a5c），表格表头为浅蓝色（#e8f0f8），NOSI提示为琥珀色。

ExternalHyperlink

格式：

论文引用：
```
https://consensus.app/papers/...
```
FOA链接：
```
https://grants.nih.gov/grants/guide/...
```

RePORTER项目：

https://reporter.nih.gov/project-details/<id>

Mandatory Program Officer Recommendation

强制要求的项目官员推荐

Always include in Section 7:

Recommended next step: contact program officer at {top institute}. Find their staff page at https://www.nih.gov/institutes-nih/list-nih-institutes-centers-offices → {institute} → Program Officers. Prepare: 1-page specific aims + your CV + 3 specific questions about fit. Email subject: "Pre-application inquiry: <topic>".

This is the single most valuable advice for any applicant. Never skip.

第7节必须包含：

推荐下一步行动：联系{排名第一的研究所}的项目官员。 在https://www.nih.gov/institutes-nih/list-nih-institutes-centers-offices → {研究所} → Program Officers查找其个人页面。准备：1页特定目标说明+你的简历+3个关于适配性的具体问题。邮件主题："Pre-application inquiry: <topic>"。

这是对任何申请者最有价值的建议。绝不跳过。

Submission Timeline (Embedded in DOCX Section 7)

提交时间表（嵌入DOCX第7节）

Mechanism	Standard receipt dates
R01, R21, R03	Feb 5, Jun 5, Oct 5
K awards (K01, K08, K23, K99)	Feb 12, Jun 12, Oct 12
R34, R61/R33	Feb 16, Jun 16, Oct 16
F31, F32	Apr 8, Aug 8, Dec 8

机制类型	标准接收日期
R01, R21, R03	2月5日、6月5日、10月5日
K类资助（K01, K08, K23, K99）	2月12日、6月12日、10月12日
R34, R61/R33	2月16日、6月16日、10月16日
F31, F32	4月8日、8月8日、12月8日

Phase 4: Deliver

第四阶段：交付

Save DOCX to

<output-dir>/grants_<topic-slug>_<YYYY-MM-DD>.docx

Chat summary: file path + audit counts + plan tier + verdict on institute targets

Validate:

python scripts/office/validate.py <docx>

将DOCX保存至

<output-dir>/grants_<topic-slug>_<YYYY-MM-DD>.docx

聊天摘要：文件路径+审计统计+计划层级+目标研究所评估结果

验证：

python scripts/office/validate.py <docx>

Tooling

工具列表

Script	Role
`scripts/citation_tracker.py`	Three-count audit (Consensus sent/shown/cited + RePORTER projects/cited) at `~/.grants_sessions/<session>.json`
`scripts/fiscal_year_calculator.py`	Current FY + 3-prior window. Computed at runtime, never hardcoded.
`scripts/mechanism_matcher.py`	Career stage × scope × prelim → mechanism recommendation shortlist

脚本	作用
`scripts/citation_tracker.py`	三计数审计（Consensus发送/展示/引用数 + RePORTER项目/引用数），存储于 `~/.grants_sessions/<session>.json`
`scripts/fiscal_year_calculator.py`	计算当前财年+前3个财年窗口。实时计算，绝不硬编码。
`scripts/mechanism_matcher.py`	职业阶段×范围×初步数据 → 机制推荐候选列表

References

参考文献

```
references/nih_mechanism_matching.md
```
— career stage × scope × prelim → mechanism canon (7+ sources)
```
references/reporter_post_patterns.md
```
— RePORTER curl POST templates + plan-tier detection (7+ sources)
```
references/docx_9_sections.md
```
— 9-section .docx spec + technical requirements (7+ sources)

```
references/nih_mechanism_matching.md
```
— 职业阶段×范围×初步数据 → 机制标准（7+来源）
```
references/reporter_post_patterns.md
```
— RePORTER curl POST模板+计划层级检测（7+来源）
```
references/docx_9_sections.md
```
— 9章节.docx规范+技术要求（7+来源）

Error Handling

错误处理

Failure	Behavior
Consensus rate-limit hit	Wait 3s, retry once, log; if still failing, alert researcher
Consensus returns 0 for a facet	Surface explicitly; never fill with training knowledge
Consensus plan-tier cap detected	Log tier, note in audit, surface to researcher
RePORTER POST returns error	Retry once after 3s; if still failing, log and continue
RePORTER returns <5 on narrow	Document; broad OR should compensate; surface low count
NOSI fetch fails	Log `[NOSI {n} — fetch failed]` , continue
3 consecutive tool failures	Stop, alert researcher with what's missing
DOCX generation fails	Save raw data as JSON fallback so researcher doesn't lose work

失败类型	处理方式
Consensus触发速率限制	等待3秒，重试一次，记录日志；若仍失败，提醒研究者
Consensus某维度返回0结果	明确说明；绝不使用训练数据填充
检测到Consensus计划层级限制	记录层级，在审计中说明，告知研究者
RePORTER POST请求返回错误	3秒后重试一次；若仍失败，记录日志并继续执行
RePORTER精确搜索返回结果<5	记录文档；宽泛OR搜索应弥补；告知结果数量较少
NOSI获取失败	记录 `[NOSI {n} — 获取失败]` ，继续执行
连续3次工具调用失败	停止操作，提醒研究者缺失内容
DOCX生成失败	将原始数据保存为JSON作为备选，避免研究者丢失工作成果

Anti-Patterns To Reject

需拒绝的反模式

Parallelizing Consensus calls (will hit rate limit)
Using
```
web_fetch
```
for RePORTER (POST-only —
```
web_fetch
```
is GET)
Hardcoded fiscal year values
Mechanism recommendations based on career stage alone (must consider scope too)
Silently filling thin facet results with training knowledge
Skipping the audit log
Skipping the program officer recommendation
Conflating "papers found" with "papers shown" with "papers cited"
Fabricating NOSI details when fetch fails

Version: 1.0.0 Source spec:

megaprompts/08-grants-megaprompt.md

Build pattern: Path B (direct conversion). Research-pack sibling of pulse + litreview.

并行执行Consensus调用（会触发速率限制）
使用
```
web_fetch
```
访问RePORTER（仅支持POST —
```
web_fetch
```
为GET）
硬编码财年值
仅基于职业阶段推荐机制（必须同时考虑范围）
使用训练数据静默填充维度结果不足的情况
跳过审计日志
跳过项目官员推荐
混淆“找到的论文数”与“展示的论文数”与“引用的论文数”
NOSI获取失败时编造细节

版本： 1.0.0 来源规范：

megaprompts/08-grants-megaprompt.md

构建模式： Path B（直接转换）。与pulse、litreview同属Research-pack系列。