Skip to content

Self-evolving Context Database for AI Agents. Unify Agent Memory, Knowledge RAG and Skills.

perf(retrieval): align default intent planner with compact v4 prompt (#6039)

Zayn Jarvisb8ec9d8210ab
diff --git a/openviking/prompts/templates/retrieval/intent_analysis.yaml b/openviking/prompts/templates/retrieval/intent_analysis.yaml
index 15b055ceb..fb5a5288a 100644
--- a/openviking/prompts/templates/retrieval/intent_analysis.yaml
+++ b/openviking/prompts/templates/retrieval/intent_analysis.yaml
@@ -1,8 +1,8 @@
metadata:
id: "retrieval.intent_analysis"
name: "Intent Analysis"
- description: "Analyze session context to generate query plans for different context types"
- version: "2.1.0"
+ description: "Generate compact query plans for different context types"
+ version: "2.2.0"
language: "en"
category: "retrieval"
@@ -36,7 +36,7 @@ variables:
required: false
template: |
- You are OpenViking's context query planner, responsible for analyzing task context gaps and generating queries.
+ You are OpenViking's context query planner. Given the session context and the current message, decide what context information is missing and emit retrieval queries to fill the gap.
## Session Context
@@ -57,119 +57,35 @@ template: |
**Target Directory Abstract**: {{ target_abstract }}
{% endif %}
- **Important**: You can only generate `{{ context_type }}` type queries, do not generate other types.
+ Only emit `{{ context_type }}` queries; do not generate other types.
{% endif %}
- ## Your Task
+ External information takes priority over built-in knowledge - actively query for any missing context.
- Analyze the current task, identify context gaps, and generate queries to fill in the required information.
+ ## Context Types
- **Core Principle**: OpenViking's external information takes priority over built-in knowledge, actively query external context.
+ - `skill` - executable capability (tool, function, API, automation). Query style: imperative starting with a verb.
+ - `resource` - knowledge artifact (doc, spec, guide, code, configuration). Query style: noun phrase.
+ - `memory` - user preference or agent execution experience.
- ## Context Types and Query Styles
+ ## Procedure
- OpenViking supports the following context types, **each type has a different query style**:
+ 1. Classify the task: operational tasks typically need skill+resource+memory; informational tasks typically need resource+memory; conversational small talk needs no query.
+ 2. Skip any context type already covered explicitly in the conversation.
+ 3. For each needed type, emit 1-5 concise retrievable queries with `priority` from 1 to 5.
- ### 1. skill (Execution Capability)
-
- **Purpose**: Executable tools, functions, APIs, automation scripts
-
- **When to Query**:
- - Task contains action verbs (create, generate, write, build, analyze, process)
- - Need to perform specific operations
-
- ### 2. resource (Knowledge Resources)
-
- **Purpose**: Documents, specifications, guides, code, configurations, and other structured knowledge
-
- **When to Query**:
- - Need reference materials, templates, specifications
- - Need to understand knowledge, concepts, definitions
-
- ### 3. memory (User/Agent Memory)
-
- **Purpose**: User personalization information or Agent execution experience
-
- **When to Query**:
- - Need personalized customization (user memory)
- - Need to learn from historical experience (agent memory)
-
- ## Analysis Method
-
- ### Step 1: Identify Task Type
-
- **Operational Tasks** (containing actions):
- - Characteristics: Verbs like create, generate, write, build, transform, calculate, analyze, process
- - Typical context combination: `skill + resource + memory`
-
- **Informational Tasks** (acquiring knowledge):
- - Characteristics: What is, how to understand, why, concept explanation, etc.
- - Typical context combination: `resource + memory`
-
- **Conversational Tasks** (small talk):
- - Characteristics: Greetings, small talk, confirmation of understanding, etc.
- - Usually no query needed
-
- ### Step 2: Check Context Coverage
-
- Analyze whether the session context (summary + recent conversation) already contains the information needed to complete the task:
-
- - **Fully covered**: Skip queries for that type
- - **Partially covered**: Generate supplementary queries
- - **Not covered**: Generate complete queries
-
- **Note**: Only skip information that has been **explicitly and in detail** discussed in the context.
-
- ### Step 3: Generate Queries
-
- **Important Principles**:
-
- 1. **Don't over-transform**:
- - ❌ Don't convert "Create XX" to "XX format/specification"
-
- 2. **Multi-type combination**:
- - A task may require multiple context types
- - Operational tasks typically need: skill (execution) + resource (reference) + memory (preference/experience)
-
- 3. **Multiple queries per type**:
- - Can generate multiple queries for the same type
- - Maximum 5 queries
-
- 4. **Queries should be concise and specific**:
- - Queries should be short, specific, and retrievable
- - Avoid lengthy descriptions
+ ## Output Format
- 5. **Priority setting**:
- - 1 = Highest priority (core requirement)
- - 3 = Medium priority (helpful)
- - 5 = Lowest priority (optional)
+ Output a single JSON object with exactly one top-level key:
- 6. **Query Style** (optimize for vector / semantic retrieval):
- - Queries are embedded and matched against indexed content by **semantic similarity**. Write each query so its embedding lands close to the target content — not necessarily a verbatim fragment, any phrasing that captures the same meaning works.
- - **Declarative, not interrogative**: state the information need as a noun/verb phrase rather than a question. Drop question framings ("what / who / when / how is ...").
- - **One information need per query**: each query targets one retrievable fact, relation, comparison, event, or procedure. Do not pile unrelated intents into one query.
- - **Self-contained**: resolve pronouns and references using the session context; the retriever only sees the query string.
- - **Concept-dense and natural**: use a grammatical, well-formed phrase carrying the key entities, attributes, and qualifiers. Avoid both bare single keywords and telegraphic word-salad.
- - **No retrieval-meta words**: exclude words describing the act of retrieval or generic containers ("find", "search", "records", "information about", "content", "details", etc.) — they do not appear in the target content and only dilute the embedding.
- - **Keep discriminative specifics**: preserve names, dates, places, and domain terms from the task — they anchor the embedding to the right content.
+ - `queries`: array of objects with:
+ - `query`: actual query text
+ - `context_type`: one of `skill`, `resource`, `memory`
+ - `priority`: integer from 1 to 5
- ## Output Format
+ If no query is needed, set `queries` to an empty array `[]`.
- ```json
- {
- "reasoning": "1. Task type (operational/informational/conversational); 2. What context is needed (skill/resource/memory); 3. What is already in context; 4. What is missing and needs to be queried",
- "queries": [
- {
- "query": "Specific query text (following the style of the corresponding type)",
- "context_type": "skill|resource|memory",
- "intent": "Purpose of the query",
- "priority": 1-5
- }
- ]
- }
- ```
-
- Please output JSON:
+ Output the JSON object directly. Do not wrap it in markdown code fences.
llm_config:
temperature: 0.0
diff --git a/tests/prompts/test_retrieval_intent_prompt_contract.py b/tests/prompts/test_retrieval_intent_prompt_contract.py
index 474389464..2f54462ba 100644
--- a/tests/prompts/test_retrieval_intent_prompt_contract.py
+++ b/tests/prompts/test_retrieval_intent_prompt_contract.py
@@ -9,11 +9,12 @@ from openviking.prompts.manager import PromptManager
@pytest.mark.parametrize(
"prompt_id",
[
+ "retrieval.intent_analysis",
"retrieval.ov_intent_analysis_sft_v4",
"retrieval.ov_intent_analysis_sft_v7",
],
)
-def test_sft_intent_prompts_use_compact_queries_contract(prompt_id: str):
+def test_intent_prompts_use_compact_queries_contract(prompt_id: str):
manager = PromptManager(templates_dir=PromptManager._get_bundled_templates_dir())
rendered = manager.render(
prompt_id,
@@ -34,3 +35,18 @@ def test_sft_intent_prompts_use_compact_queries_contract(prompt_id: str):
assert "priority" in output_contract
assert '"reasoning"' not in output_contract
assert '"intent"' not in output_contract
+
+
+def test_default_intent_prompt_matches_v4_contract():
+ manager = PromptManager(templates_dir=PromptManager._get_bundled_templates_dir())
+ variables = {
+ "compression_summary": "The user is working on OpenViking retrieval.",
+ "recent_messages": "[user]: Improve semantic search.",
+ "current_message": "Use the style we discussed earlier.",
+ "context_type": "resource",
+ "target_abstract": "OpenViking retrieval documentation.",
+ }
+
+ assert manager.render("retrieval.intent_analysis", variables) == manager.render(
+ "retrieval.ov_intent_analysis_sft_v4", variables
+ )
diff --git a/tests/retrieve/test_intent_analyzer_query_planner.py b/tests/retrieve/test_intent_analyzer_query_planner.py
index b805bbbb2..da8ce03f0 100644
--- a/tests/retrieve/test_intent_analyzer_query_planner.py
+++ b/tests/retrieve/test_intent_analyzer_query_planner.py
@@ -153,6 +153,35 @@ async def test_intent_analyzer_handles_bare_array_response(monkeypatch):
assert result.reasoning == ""
+async def test_intent_analyzer_handles_compact_object_response(monkeypatch):
+ planner = RecordingModel(
+ """{
+ "queries": [
+ {
+ "query": "compact query",
+ "context_type": "resource",
+ "priority": 2
+ }
+ ]
+ }"""
+ )
+ monkeypatch.setattr(intent_module, "render_prompt", lambda prompt_id, variables: "prompt")
+
+ result = await IntentAnalyzer(query_planner=planner).analyze(
+ compression_summary="",
+ messages=[],
+ current_message="find the relevant documentation",
+ )
+
+ assert len(result.queries) == 1
+ assert result.queries[0].query == "compact query"
+ assert result.queries[0].context_type.value == "resource"
+ assert result.queries[0].intent == ""
+ assert result.queries[0].priority == 2
+ assert result.reasoning == ""
+
+
@pytest.mark.asyncio
async def test_intent_analyzer_uses_model_specific_prompt(monkeypatch):
planner = RecordingModel(