Knowledge Locator
Model and maintain spatial, semantic, and temporal indices for memory-palace agent knowledge so concepts can be located by palace path, keywords, or access patterns.
Overview
knowledge-locator is an agent skill for the Build phase that defines spatial, semantic, and temporal index schemas for memory-palace agent knowledge retrieval.
Install
npx skills add https://github.com/athola/claude-night-market --skill knowledge-locatorWhat is this skill?
- Documents primary index JSON schema: spatial coordinates from palace through area plus concept_data and sensory_signatur
- Three index types: spatial (hierarchical O(log n)), semantic (clusters + relationship graph), temporal (timeline, access
- Supports wildcard matching at any spatial hierarchy level
- Covers index maintenance including rebuild operations (per index-structure spec)
- Tagged advanced complexity with ~350 estimated tokens in frontmatter metadata
- Three index types: spatial, semantic, temporal
- Spatial hierarchical lookup described as O(log n) traversal
Adoption & trust: 1 installs on skills.sh; 304 GitHub stars; 3/3 security scanners passed (skills.sh audits); trending (+100% hot-view momentum).
What problem does it solve?
Your agent stores palace-shaped knowledge but lacks a consistent multi-index schema for location, meaning, and freshness queries.
Who is it for?
Builders implementing custom agent memory or memory-palace architectures who need a shared index contract across skills.
Skip if: Teams wanting turnkey RAG over PDFs, simple note search, or launch/growth analytics without building agent memory infrastructure.
When should I use this skill?
Implementing or extending knowledge-locator / index-structure dependencies when designing agent memory indexing.
What do I get? / Deliverables
You implement aligned spatial, semantic, and temporal indices with documented maintenance and lookup characteristics for downstream locator logic.
- Index schema definitions for spatial, semantic, and temporal stores
- Maintenance guidance for index rebuild and decay handling
Recommended Skills
Journey fit
Indexing schemas and maintenance rules are build-time agent infrastructure—the skill defines how agent memory is stored and queried, not go-to-market or ops monitoring. Agent-tooling is the right shelf for data structures (spatial_index, semantic_index, temporal_index) that power multi-modal retrieval in custom agent stacks.
How it compares
Reference data-structure spec for agent memory—not a hosted vector DB integration or a documentation co-authoring workflow.
Common Questions / FAQ
Who is knowledge-locator for?
Advanced solo builders and agent authors wiring memory-palace or multi-modal knowledge systems who need index schemas and maintenance rules.
When should I use knowledge-locator?
During build while designing agent-tooling storage and retrieval—when defining how palace/district/room paths, keywords, and access decay map to query APIs.
Is knowledge-locator safe to install?
Treat it as documentation-heavy agent guidance; review the Security Audits panel on this Prism page before enabling tools in your agent runtime.
SKILL.md
READMESKILL.md - Knowledge Locator
# Index Structure The Knowledge Locator maintains multiple indices for efficient multi-modal search. ## Primary Index Schema ```json { "spatial_index": { "coordinates": { "palace": "string", "district": "string", "building": "string", "room": "string", "area": "string" }, "concept_data": { "primary_concept": "string", "keywords": ["list"], "sensory_signature": "object", "associations": ["related_concepts"], "access_frequency": "number", "last_accessed": "timestamp" } }, "semantic_index": { "concept_clusters": {}, "relationship_graph": {}, "context_mappings": {} }, "temporal_index": { "creation_timeline": {}, "access_patterns": {}, "decay_rates": {} } } ``` ## Index Types ### Spatial Index - Hierarchical path-based lookup - O(log n) traversal for location queries - Supports wildcard matching at any level ### Semantic Index - Keyword-to-concept mapping - Concept clustering for similarity search - Relationship graph for association traversal ### Temporal Index - Creation timeline for history queries - Access frequency for hot/cold optimization - Decay rates for staleness detection ## Index Maintenance - **Rebuild**: Full reindex when structure changes significantly - **Incremental**: Update affected entries on concept changes - **Compact**: Remove deleted entries and optimize storage --- name: search-strategies description: Multi-modal search strategies and optimization techniques category: algorithms tags: [search, algorithms, optimization] dependencies: [knowledge-locator] complexity: intermediate estimated_tokens: 300 --- # Search Strategies Effective search combines multiple strategies based on query type and context. ## Strategy Selection | Query Type | Primary Strategy | Secondary | |-----|-----|-----| | Exact path | Spatial lookup | - | | Keyword | Semantic search | Fuzzy match | | Partial info | Multi-modal fusion | Associative walk | | Discovery | Graph traversal | Random walk | ## Strategy Implementations ### Spatial Lookup 1. Parse path components 2. Traverse hierarchy 3. Return exact match or nearest ancestors ### Semantic Search 1. Extract query keywords 2. Match against semantic index 3. Rank by relevance score 4. Return top-k results ### Fuzzy Matching 1. Apply edit distance tolerance 2. Check phonetic similarity 3. Expand query with synonyms 4. Merge and rank results ### Associative Walk 1. Start from known concept 2. Follow association edges 3. Score by path relevance 4. Return connected concepts ## Optimization Techniques - **Query caching** - Cache frequent queries - **Index partitioning** - Split by palace for parallelism - **Hot path optimization** - Preload frequently accessed paths - **Lazy loading** - Load deep indices on demand --- name: knowledge-locator description: Searches and navigates stored knowledge in memory palaces. Use when looking for previously stored information or cross-referencing concepts across palaces. alwaysApply: false category: navigation tags: - retrieval - search - indexing - recall - spatial-memory - pr-review dependencies: - memory-palace-architect - review-chamber scripts: - palace_manager.py usage_patterns: - search - cross-reference - discovery - review-search complexity: intermediate model_hint: standard estimated_tokens: 500 --- ## Table of Contents - [What It Is](#what-it-is) - [Quick Start](#quick-start) - [Search Palaces](#search-palaces) - [List All Palaces](#list-all-palaces) - [When to Use](#when-to-use) - [Search Modalities](#search-modalities) - [Core Workflow](#core-workflow) - [Target Metrics](#target-metrics) - [Detailed Resources](#detailed-resources) - [PR Review Search](#pr-review