메뉴
Recall memories from World Weaver using multi-strategy retrieval
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
| name | ww-recall |
| description | Recall memories from World Weaver using multi-strategy retrieval |
| version | 1.0.0 |
| allowed-tools | ["Bash","Read"] |
Multi-strategy memory retrieval from World Weaver's tripartite memory system.
Search and retrieve relevant memories from:
Invoke this skill when:
mcp__ww-memory__recall_episodes - Search episodic memory
mcp__ww-memory__semantic_recall - Search knowledge graph
mcp__ww-memory__spread_activation - Graph traversal from seed
mcp__ww-memory__recall_skill - Find applicable skills
mcp__ww-memory__query_at_time - Point-in-time queries
Best for: General queries, finding similar content
mcp__ww-memory__recall_episodes(
query: "your search query",
limit: 10,
session_filter: null # All sessions, or specific session_id
)
Returns episodes ranked by:
Best for: "What did we do last week?", time-bounded queries
mcp__ww-memory__recall_episodes(
query: "project work",
limit: 10,
time_filter: {
after: "2025-11-20T00:00:00",
before: "2025-11-27T23:59:59"
}
)
Best for: Finding entities and their relationships
mcp__ww-memory__semantic_recall(
query: "concept or entity name",
limit: 10,
include_relationships: true
)
Best for: Exploring connections from a known starting point
mcp__ww-memory__spread_activation(
seed_entities: ["Entity A", "Entity B"],
max_depth: 2,
min_weight: 0.3
)
Returns entities connected to seeds, weighted by relationship strength.
Best for: "How do I...?", finding applicable procedures
mcp__ww-memory__recall_skill(
query: "how to run tests",
limit: 5,
check_preconditions: true,
context: {
project: "ww",
working_directory: "/home/aaron/ww"
}
)
Best for: "What did we know at time X?"
mcp__ww-memory__query_at_time(
query: "project status",
as_of: "2025-11-15T12:00:00"
)
For comprehensive recall, combine strategies:
Determine what the user is looking for:
Run appropriate strategies in parallel:
episodes = mcp__ww-memory__recall_episodes(query, limit=5)
entities = mcp__ww-memory__semantic_recall(query, limit=5)
skills = mcp__ww-memory__recall_skill(query, limit=3)
Combine results, removing duplicates by ID.
Re-rank by overall relevance:
Present as structured context:
## Memory Context
### Recent Episodes (3 found)
1. **[2025-11-27]** Fixed batch query bug in Neo4j store
- Outcome: success
- Relevance: 0.92
2. **[2025-11-26]** Implemented session isolation tests
- Outcome: success
- Relevance: 0.85
### Related Knowledge (2 entities)
1. **Neo4j** (CONCEPT)
- Graph database used for relationships
- Connected to: Qdrant, Cypher, World Weaver
2. **Batch Queries** (CONCEPT)
- Optimization pattern for N+1 elimination
### Applicable Skills (1 found)
1. **run-ww-tests**
- Run World Weaver test suite
- Preconditions: In WW directory, venv exists
| Parameter | Type | Description |
|---|---|---|
query | string | Search query text |
limit | int | Max results (default: 10) |
session_filter | string | Filter by session ID |
time_filter | object | After/before timestamps |
| Parameter | Type | Description |
|---|---|---|
outcome_filter | string | success/failure/partial/neutral |
min_valence | float | Minimum importance (0-1) |
| Parameter | Type | Description |
|---|---|---|
entity_type | string | CONCEPT/PERSON/PLACE/etc |
include_relationships | bool | Include connected entities |
| Parameter | Type | Description |
|---|---|---|
check_preconditions | bool | Verify preconditions met |
min_success_rate | float | Minimum skill success rate |
User: "What have we worked on recently?"
Action:
mcp__ww-memory__recall_episodes(
query="recent work projects tasks",
limit=10,
time_filter={after: "7 days ago"}
)
Output:
Found 8 recent episodes:
1. [Nov 27] Plugin architecture planning for World Weaver
2. [Nov 27] Fixed batch query bugs, session isolation
3. [Nov 26] Implemented HDBSCAN memory clustering
...
User: "How do I run the integration tests?"
Action:
mcp__ww-memory__recall_skill(
query="run integration tests",
limit=3,
check_preconditions=true
)
Output:
Found skill: run-ww-tests
Steps:
1. source .venv/bin/activate
2. pytest tests/integration/ -v -m integration
3. Check output for failures
User: "What do we know about Neo4j?"
Action:
1. mcp__ww-memory__semantic_recall(query="Neo4j", limit=1)
2. mcp__ww-memory__spread_activation(seed_entities=["Neo4j"], max_depth=2)
Output:
## Neo4j (CONCEPT)
Graph database for relationship storage.
### Connected Entities:
- Cypher (query language) - strength: 0.9
- Qdrant (co-storage) - strength: 0.8
- World Weaver (uses) - strength: 0.95
- Batch Queries (optimization) - strength: 0.7
User: "What did we do last week on the PhD project?"
Action:
mcp__ww-memory__recall_episodes(
query="PhD dissertation xai",
limit=10,
time_filter={
after: "2025-11-20",
before: "2025-11-27"
}
)
Output:
Found 4 episodes from Nov 20-27:
1. [Nov 25] Committee meeting preparation
2. [Nov 23] VGGFace2 dataset removal analysis
3. [Nov 21] Article C IEEE T-BIOM submission
4. [Nov 20] Experiment 6.1 planning
When presenting recall results:
If recall fails:
Comprehensive SessionEnd hook that syncs and synthesizes all daily completions, notes, todos, and conversation data into unified daily notes and knowledge bases.
Unified worklog system. Synthesizes session work into daily markdown files with clean bullet points. Supports weekly rollups for Slack #progress sharing. Replaces fragmented checkpoint/session-state/weekly-notes systems.
Comprehensive pre-submission review orchestrator. Coordinates methodology checks, bias detection, AI text detection, statistical validation, and reporting guideline compliance for journal submissions.
Kymera Systems brand design - Jarvis-inspired HUD aesthetic with dark and light modes. Use for branded UI, dashboards, presentations, and technical artifacts.
Deep analysis workflows for World Weaver memory systems, code, and architecture
Architecture documentation generation and maintenance for World Weaver