菜单
Analyze BRD and existing corpus to identify knowledge gaps, then generate prioritized research tasks for filling those gaps
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Outline creation patterns for fiction and nonfiction book structures, including chapter breakdowns, scene planning, and narrative arc mapping.
Grounded prose composition workflow including pre-write queries, consistency checks, and post-write entity extraction for database-backed writing.
Run editing passes over completed sections for voice consistency, timeline verification, contradiction flagging, and citation coverage
Load research corpus into the database by processing files, directories, or URLs through semantic chunking, embedding generation, entity extraction, and relationship building
Generate writing tasks from BRD structure and corpus, creating chapter/section tasks with dependencies, pre-write queries, and consistency checks
Execute ad-hoc database queries to explore the book's knowledge base using natural language questions that translate to semantic search, graph traversal, timeline queries, or hybrid queries
| name | bookstrap-plan-research |
| description | Analyze BRD and existing corpus to identify knowledge gaps, then generate prioritized research tasks for filling those gaps |
| disable-model-invocation | true |
| allowed-tools | Bash, Read |
Analyze the Book Requirements Document (BRD) against the existing research corpus to identify knowledge gaps and generate prioritized research tasks.
Compare what the book requires (from BRD) with what knowledge currently exists in the database (from prior ingestion). Generate a prioritized list of research tasks that will fill the identified gaps before writing begins.
This command delegates the analysis work to the corpus-analyst agent.
None. This command reads from the database:
Query the database to retrieve the Book Requirements Document:
# Query BRD from database
surreal sql --conn http://localhost:2665 \
--user root --pass root \
--ns bookstrap --db <database-name> \
--query "SELECT * FROM brd ORDER BY version DESC LIMIT 1;"
Extract key requirements:
Analyze what knowledge already exists in the database:
-- Coverage by chapter/topic
SELECT chapter, count() as sources
FROM section->cites->source
GROUP BY chapter;
-- Entity coverage
SELECT name, count() as mentions
FROM concept<-related_to<-source
GROUP BY name;
-- Existing knowledge gaps (unresolved)
SELECT * FROM knowledge_gap
WHERE resolved = false
ORDER BY created_at;
-- Source distribution by type
SELECT source_type, count() as total
FROM source
GROUP BY source_type;
-- Timeline coverage
SELECT sequence, count() as events
FROM event
GROUP BY sequence;
For each chapter/topic/concept in the BRD:
Create research tasks with priorities based on:
High Priority:
Medium Priority:
Low Priority:
For each research task, create a database record:
CREATE knowledge_gap SET
question = $research_question,
context = $why_needed,
priority = $priority_level,
chapter_blocking = $chapters_affected,
resolved = false,
created_at = time::now()
;
Invoke the corpus-analyst agent to perform the detailed gap analysis:
# Load corpus-analyst agent with context
# Agent will:
# 1. Read BRD from database
# 2. Execute coverage queries
# 3. Perform gap analysis
# 4. Generate prioritized research tasks
# 5. Store tasks in database
# 6. Generate report
The corpus-analyst agent has read-only database access and uses the corpus-analysis and surrealdb skills.
Report generated research tasks to the user:
KNOWLEDGE GAP ANALYSIS
======================
BRD: <Book Title>
Corpus: <N> sources, <M> entities
COVERAGE ANALYSIS
-----------------
High Coverage:
✓ Chapter 1: Introduction (8 sources, 15 entities)
✓ Concept: SOE Training (12 sources)
Medium Coverage:
~ Chapter 3: Lyon Operation (3 sources, 8 entities)
~ Location: Lyon Safehouse (2 sources)
Low/No Coverage:
✗ Chapter 5: Wireless Protocols (0 sources)
✗ Event: Gestapo Raid 1943 (0 sources)
✗ Character: Marie Dubois (mentioned in BRD, not in corpus)
RESEARCH TASKS GENERATED: 12
=============================
HIGH PRIORITY (4 tasks)
-----------------------
1. SOE wireless operator training protocols 1942-1943
Why: Core to protagonist's role, zero sources
Blocks: Chapters 2, 4, 6
2. Lyon Resistance network structure and key figures
Why: Primary setting, only 2 tangential sources
Blocks: Chapters 3, 5, 7
3. Gestapo counter-intelligence methods in Lyon 1943
Why: Antagonist strategy, no documentation
Blocks: Chapters 5, 8
4. Daily life details in occupied Lyon (food, curfews, rationing)
Why: Atmosphere and authenticity, minimal coverage
Blocks: Chapters 3, 4, 5, 7
MEDIUM PRIORITY (5 tasks)
-------------------------
5. SOE Beaulieu training facility layout and procedures
6. Wireless set types used by F Section agents
7. Lyon geography and landmarks 1943
8. French Resistance communication codes
9. Black market operations in wartime Lyon
LOW PRIORITY (3 tasks)
----------------------
10. Fashion and clothing styles 1943 France
11. Popular culture and entertainment during occupation
12. Weather patterns in Lyon region
NEXT STEPS
----------
Run /bookstrap-research to execute these tasks autonomously.
Use /bookstrap-status to monitor progress.
Calculate and report:
This command is a thin wrapper that:
corpus-analyst agentThe actual analysis logic lives in the corpus-analyst agent to keep concerns separated.
The corpus-analyst uses multiple approaches:
Re-running /bookstrap-plan-research should:
Gap detection thresholds can be configured in bookstrap.config.json:
{
"gap_detection": {
"min_sources_per_chapter": 5,
"min_sources_per_concept": 2,
"high_priority_threshold": 0,
"medium_priority_threshold": 2
}
}
| Error | Recovery |
|---|---|
| BRD not found | Abort with message to run /bookstrap-init first |
| Database connection failed | Abort with message to start SurrealDB |
| No corpus exists | Warning that all BRD requirements are gaps, generate exhaustive research plan |
| Corpus-analyst agent error | Display error, suggest re-running or checking logs |
Before running /bookstrap-plan-research:
/bookstrap-init must have been run/bookstrap-ingest provides baseline, but can run on empty corpus/bookstrap-init - Create BRD/bookstrap-ingest - Load initial corpus/bookstrap-research - Execute research tasks generated by this command/bookstrap-status - View gap resolution progress/bookstrap-plan-write - Generate writing tasks after gaps filled| Agent | Role |
|---|---|
corpus-analyst | Performs gap analysis, generates research tasks |
| Skill | Purpose |
|---|---|
corpus-analysis/ | Gap detection strategies and coverage analysis |
surrealdb/ | Database query patterns for coverage analysis |
# After initial ingestion, identify what's missing
/bookstrap-plan-research
# Re-run after additional ingestion to update gap list
/bookstrap-ingest ./more-sources/
/bookstrap-plan-research
# Check gaps before starting to write
/bookstrap-plan-research
/bookstrap-status
This command is part of the research cycle:
init → ingest → plan-research → research
↑ │
└──────────────┘
(iterate until
gaps resolved)
Once gaps are identified, /bookstrap-research executes the tasks autonomously. If new gaps are discovered during writing, re-run this command to generate new research tasks.