/garden-explore

Infrastructure L2+ Production low usage

/garden-explore

Overview

/garden-explore is a production command in the Infrastructure category of the Prismatic Platform that provides interactive exploration of GARDEN (Growing Archive of Reusable Development and Engineering Nuggets) repositories for patterns, knowledge, and reusable components. The GARDEN ecosystem comprises 116 repositories spanning over 20 years of software engineering practice, containing 3,050+ files, 55+ documented patterns, and implementations across multiple languages and frameworks. The /garden-explore command is the primary discovery tool for navigating this legacy knowledge base.

The command operates under the L2+ authority level and is executed by the garden-cultivator agent. It is part of the platform's 216-command slash command registry, built on the AIAD (Autonomous Intelligence Agent Design) standard. The garden-cultivator agent maintains a comprehensive index of all GARDEN repositories, their contents, and the patterns they contain, enabling rapid discovery of relevant knowledge regardless of the repository's age or technology stack.

Unlike /garden-search, which performs text-based pattern matching, /garden-explore provides a structured, navigable view of the GARDEN ecosystem organized by repository tier, technology domain, pattern category, and relevance to the current platform context. This makes it suitable for open-ended exploration where the operator does not yet know what they are looking for -- a common scenario when investigating architectural options, seeking precedent for a design decision, or mining historical implementations for reusable components.

The GARDEN ecosystem is organized into five tiers based on maturity and relevance: T1 Production (active, production-quality repositories like sig and prismatic), T2 Active (maintained repositories under active development), T3 Libraries (stable utility libraries), T4 Archive (historical repositories with extractable value), and T5 R&D (experimental and research repositories). The /garden-explore command respects this tiering system, prioritizing results from higher tiers while still providing access to the full archive.

Architecture

The /garden-explore command is built on a three-layer architecture: index, navigator, and presenter.

GARDEN Repos (116) --> Indexer --> Knowledge Graph --> Navigator --> Presenter
      |                   |              |                |            |
  Git Submodules     Pattern         KuzuDB/ETS       Interactive    Formatted
  in garden/        Extraction       Storage          Traversal      Output

Component Layers

LayerComponentResponsibility
IndexGarden.IndexerScans repositories, extracts metadata, builds knowledge graph
StorageGarden.KnowledgeGraphStores repository relationships, patterns, and cross-references
NavigatorGarden.NavigatorProvides structured traversal of the knowledge graph
PresenterGarden.PresenterFormats exploration results for display

Repository Classification

TierCountExamplesQuality Level
T1 Production5sig (OSINT), prismatic (AI)Production-grade, actively maintained
T2 Active8kuzu-ex, crisstal, code-weaverUnder active development
T3 Libraries15simple_geocoder, job-processorStable, minimal maintenance
T4 Archive70+prismatic-legacy (1,302 files)Historical, extractable value
T5 R&D18comtesse (Rust), prismatic-scrapperExperimental, research

Usage

Basic Usage

# Explore the entire GARDEN ecosystem (top-level overview)
/garden-explore

# Explore a specific repository
/garden-explore sig

# Explore repositories by tier
/garden-explore --tier t1

# Explore repositories by technology
/garden-explore --tech elixir

Pattern Discovery

# Explore pattern categories across all repositories
/garden-explore --patterns

# Find repositories containing blackboard pattern implementations
/garden-explore --pattern blackboard

# Explore OSINT-related patterns and providers
/garden-explore --domain osint

# Find repositories with GenServer implementations
/garden-explore --pattern genserver --tech elixir

Deep Exploration

# Explore a repository's internal structure
/garden-explore sig --depth 3

# Show file statistics for a repository
/garden-explore prismatic-legacy --stats

# Explore cross-repository dependencies
/garden-explore --dependencies

# Find repositories that share common patterns
/garden-explore --shared-patterns --min-repos 3

Options & Parameters

OptionTypeDefaultDescription
REPOstringnoneSpecific repository name to explore
--tierstringallFilter by tier: t1, t2, t3, t4, t5
--techstringallFilter by technology: elixir, python, rust, go, javascript
--domainstringallFilter by domain: osint, ai, storage, web, security
--patternsbooleanfalseShow pattern categories instead of repositories
--patternstringnoneFilter by specific pattern name
--depthinteger1Exploration depth for repository structure
--statsbooleanfalseShow file statistics and metrics
--dependenciesbooleanfalseShow cross-repository dependency graph
--shared-patternsbooleanfalseFind patterns shared across multiple repositories
--min-reposinteger2Minimum repository count for shared pattern filter
--formatstringtableOutput format: table, json, tree, markdown
--verbosebooleanfalseInclude detailed metadata for each result

Execution Flow

  1. Scope Resolution: Determine the exploration scope based on provided arguments -- full ecosystem, specific tier, technology filter, or individual repository.

  2. Index Verification: Check that the GARDEN index is current. If the index is stale (last updated more than 24 hours ago), trigger an incremental re-index of modified repositories.

  3. Knowledge Graph Query: Query the knowledge graph for entities matching the exploration criteria. For pattern exploration, this involves traversing pattern-repository relationships. For repository exploration, it involves loading repository metadata and structure.

  4. Tier-Based Prioritization: Results are sorted with T1 repositories first, followed by T2, T3, T4, and T5. Within each tier, results are sorted by relevance to the query.

  5. Depth Expansion: If --depth is specified, expand the results to show internal repository structure up to the requested depth level.

  6. Cross-Reference Resolution: Resolve cross-references between repositories and patterns, annotating results with related items.

  7. Presentation: Format the results according to the specified output format and display them.

Integration Points

ComponentIntegration TypeDescription
Prismatic AgentsExecutionInvoked by garden-cultivator agent
GARDEN EcosystemCorePrimary interface to the 116-repository knowledge base
/garden-extractWorkflowExploration findings feed into extraction operations
/garden-searchComplementaryText search complements structured exploration
/garden-syncPrerequisiteSync ensures repositories are current before exploration
/gardenerParentGardener provides top-level GARDEN management
KuzuDBStorageKnowledge graph stored in KuzuDB for relationship queries
TelemetryMetricsExploration events tracked for usage analysis

Best Practices

Start broad, then narrow. Begin with /garden-explore to get the ecosystem overview, then drill into specific tiers or domains. This prevents tunnel vision and may surface unexpected relevant repositories.

Use pattern exploration for architectural decisions. Before designing a new subsystem, run /garden-explore --patterns to see what patterns have been successfully applied across the GARDEN ecosystem. Historical precedent is a valuable signal for architectural decision-making.

Sync before exploring. Run /garden-sync before deep exploration to ensure that repository contents are current. Stale submodules may cause the explorer to miss recently added patterns or files.

Combine exploration with extraction. After identifying a relevant pattern or component through exploration, use /garden-extract to bring the discovered knowledge into the active platform. Exploration without extraction leaves value on the table.

Leverage tier prioritization for time-constrained work. When time is limited, restrict exploration to T1 and T2 repositories with --tier t1 or --tier t2. These tiers contain the highest-quality, most relevant implementations.

Error Handling

ErrorCauseResolution
{:error, :repo_not_found}Specified repository name does not match any GARDEN repoCheck repository name spelling; use /garden-explore without arguments to list all repos
{:error, :index_stale}GARDEN index is older than 24 hoursRun /garden-sync to update the index
{:error, :submodule_missing}Git submodule is not initializedRun git submodule update --init garden/REPO_NAME
{:error, :knowledge_graph_unavailable}KuzuDB storage is not accessibleVerify KuzuDB configuration in config/config.exs
{:error, :pattern_not_indexed}Specified pattern is not in the indexThe pattern may exist but not yet be indexed; run manual re-index

Advanced Usage

Cross-Domain Pattern Analysis

# Find patterns that appear in both OSINT and AI domains
/garden-explore --shared-patterns --domain osint,ai

# Explore how the blackboard pattern evolved across repositories
/garden-explore --pattern blackboard --format tree --verbose

# Map the technology migration path across tiers
/garden-explore --stats --format json | jq '.[] | {name, tier, primary_tech}'

Knowledge Extraction Pipeline

Combine exploration with extraction for systematic knowledge mining:

# Step 1: Explore to identify candidates
/garden-explore --pattern provider --tech elixir --tier t1,t2

# Step 2: Extract identified patterns
/garden-extract sig --pattern osint-provider --target apps/prismatic_agents/

# Step 3: Verify integration
/garden-search --verify-integration osint-provider

GARDEN Metrics and Health

# Get comprehensive GARDEN statistics
/garden-explore --stats --format markdown > garden-health-report.md

# Identify repositories with the most extractable value
/garden-explore --stats --tier t4 --format json | jq 'sort_by(-.file_count) | .[0:10]'

Doctrine Compliance

All commands operate under the NO MERCY, NO DOUBTS doctrine:

  • NO MERCY: Zero tolerance for incomplete execution or quality violations. Exploration results include only verified, indexed content -- never stale or corrupted data.
  • NO DOUBTS: Full investigation before action, evidence-based results. Exploration provides comprehensive context before extraction decisions are made, ensuring that knowledge transfer is deliberate and evidence-based.

The command supports NABLA signal plurality by surfacing multiple pattern implementations across different repositories and technology stacks, enabling operators to compare approaches and select the most appropriate solution based on diverse evidence.

  • /gardener - GARDEN legacy knowledge repository management across 116 repos
  • /garden-extract - Extract and integrate patterns from GARDEN repositories
  • /garden-search - Fast pattern search across all GARDEN reference repositories
  • /garden-sync - Synchronize GARDEN submodules to latest remote commits
  • /ollama - Local AI Ollama model management, installation and optimization
  • /propagate-pattern - Propagate successful patterns across the ecosystem

Connect & Contribute

Created by Tomáš Korcak (korczis) | Open Source under GHL