strategy-review - SKILL.md Agent Skill

name: strategy-review description: Search and analyse global health strategy documents using the knowledge graph and document corpus argument-hint: Ask a question about global health strategies (e.g. "What funding is allocated to TB?")

Strategy Review Skill

Answer questions about global health strategy documents by querying the Neo4j knowledge graph and OpenSearch document corpus.

Workflow

Step 1 — Receive the Question

The user provides their question via $ARGUMENTS.

If $ARGUMENTS is empty, ask the user what they'd like to know about the global health strategies. Suggest example questions:

"How much funding is allocated to TB elimination?"
"What does the Global Health Strategy say about malaria prevention?"
"Which countries prioritise maternal health and what interventions are recommended?"

Step 2 — Classify the Question

Before dispatching, classify the question into one of three categories:

Graph — use when the question asks about:

Relationships between entities (documents, themes, countries, indicators, funding areas)
Funding amounts, budget allocations, or percentages
Country priorities, rankings, or implementation progress
Indicator baselines, targets, or measurements
Comparisons across entities (e.g. "which theme gets the most funding?")

Text — use when the question asks about:

Specific policy details, recommendations, or interventions
Quotes or passages from strategy documents
Section content or document summaries
Contextual explanations or rationale behind strategies

Both — use when:

The question needs structural data AND textual evidence
The question is complex or multi-part
The category is ambiguous — when in doubt, use Both

Classification examples:

"How much funding is allocated to TB?" → Graph
"What are the TB treatment success rate targets?" → Graph
"Which countries support HIV programmes?" → Graph
"What does the GH Strategy say about malaria prevention?" → Text
"Summarise the TB Elimination Plan's approach to drug-resistant TB" → Text
"What gender-responsive budgeting recommendations exist?" → Text
"Which countries prioritise maternal health and what interventions are recommended?" → Both
"What is the funding for digital health and how will it be spent?" → Both
"Compare TB approaches across the Global Health Strategy and TB Elimination Plan" → Both

Step 3 — Dispatch to Sub-Agents

Based on the classification, dispatch using the Task tool.

If Graph:

Launch a single Task with subagent_type: "general-purpose" and the following prompt:

You are a Neo4j Cypher specialist querying a global health strategy knowledge graph.

Use the `mcp__neo4j__read_neo4j_cypher` tool to answer this question:

${ARGUMENTS}

## Graph Schema

Node types:
- Document: id, title, type, year, organization, region
- Theme: id, name, description, priority
- Indicator: id, name, unit, target
- Country: code, name, region, income_level
- FundingArea: id, name, budget_usd_millions, fiscal_year

Relationships:
- (Document)-[:COVERS_THEME {primary, weight}]->(Theme)
- (Theme)-[:MEASURED_BY {baseline, target, year}]->(Indicator)
- (Theme)-[:PRIORITY_IN {rank, rationale}]->(Country)
- (FundingArea)-[:ALLOCATES_TO {amount_usd_millions, percentage}]->(Theme)
- (Country)-[:SUPPORTS_THEME {implementation_status, progress_pct}]->(Theme)

Guidelines:
- Use parameterised queries where possible
- Always include LIMIT clauses
- Return specific properties, not full nodes
- Present results as structured data with context

If Text:

Launch a single Task with subagent_type: "general-purpose" and the following prompt:

You are a text search specialist querying a global health strategy document corpus.

Use the Strategy Review MCP tools to answer this question:

${ARGUMENTS}

## Available Tools

1. `mcp__strategy-review__search_documents(query, top_k=5)` — broad document-level search
2. `mcp__strategy-review__search_chunks(query, doc_id=None, top_k=5)` — granular chunk search, optionally filtered by doc_id
3. `mcp__strategy-review__get_page_image(doc_id, page_num)` — retrieve original page image

## Document Corpus

- GH_2024: Global Health Strategy 2024-2028 (Global Fund)
- TB_2025: TB Elimination Plan 2025-2030 (WHO)
- GE_2023: Gender & Health Equity Framework 2023-2028 (Global Fund)

## Search Strategy

1. Start broad with search_documents to find relevant documents
2. Drill down with search_chunks filtered by doc_id for specific passages
3. Always cite: document title, section name, and page number

Present results as structured summaries with exact quotes where relevant.

If Both:

Launch TWO Task calls in parallel (both with subagent_type: "general-purpose"), one with the Graph prompt and one with the Text prompt above. Wait for both to return before proceeding to synthesis.

Step 4 — Synthesise Results with Citations

Once the agent(s) return, synthesise the results into a single coherent response.

Output Format

## Answer

[Concise 2-3 sentence summary directly answering the question]

## Evidence

[Detailed supporting information organised by sub-topic. Every claim must have an inline citation.]

For graph-sourced data, cite as: *(Knowledge Graph: Node→Relationship→Node)*
For text-sourced data, cite as: *(Document Title, Section Name, p.XX)*

## Sources

[Numbered reference list]
1. [Document or data source with specific section/query referenced]
2. ...

Synthesis Guidelines

Lead with the direct answer — don't bury it in evidence
When both agents return data, weave graph and text results together rather than presenting them in separate blocks
If graph data and text data conflict, flag the discrepancy and present both
Use tables for comparative data (funding amounts, country rankings, indicator targets)
Quote document text exactly when it adds value — use block quotes with citation

Step 5 — Offer Follow-Up Questions

After presenting the synthesised answer, suggest 2-3 related questions the user might want to explore next. Base these on:

Adjacent themes or countries mentioned in the results
Deeper dives into data points surfaced
Cross-cutting questions that span both graph and text

Format as:

**Want to dig deeper?**
- [Related question 1]
- [Related question 2]
- [Related question 3]

Example Usage

/strategy-review What funding is allocated to maternal and newborn health?

Expected classification: Graph (funding allocation query) Expected dispatch: graph-traversal agent queries FundingArea→ALLOCATES_TO→Theme Expected output: table of funding areas with amounts, total sum, with citations to graph structure

/strategy-review What does the Gender & Health Equity Framework recommend for reducing maternal mortality?

Expected classification: Text (policy recommendation query) Expected dispatch: document-search agent searches GE_2023 chunks Expected output: summary of recommendations with quotes, section and page citations

/strategy-review Which countries prioritise TB elimination and what interventions are planned?

Expected classification: Both (country priorities from graph + intervention details from text) Expected dispatch: both agents in parallel Expected output: country priority rankings from graph woven with intervention details from text