Agent skill

get-pattern

Retrieve APPLICATION patterns (architecture, procedures, conventions) from AgentDB skills table. Use BEFORE implementing to ensure consistency.

View SKILL.md on GitHub Repository
Stars 163
Forks 31

Install this agent skill to your Project

npx add-skill https://github.com/majiayu000/claude-skill-registry/tree/main/skills/development/get-pattern

SKILL.md

Get Pattern - Retrieve Application Knowledge

What This Skill Does

Retrieves established application patterns (architecture, procedures, conventions) for the Neural Data Platform using AgentDB's semantic skill search.

Use this BEFORE implementing anything to ensure you follow project standards.

Quick Reference

bash
# Search for patterns by description
agentdb skill search "domain adapter pattern" 5

# Fallback: search reflexion episodes for past experiences
agentdb reflexion retrieve "how to add a stream" --k 5 --only-successes

# View all stored patterns
agentdb db stats

Primary Method: Skill Search

bash
agentdb skill search "<query>" <k>

Parameters

Parameter Description
<query> What you're looking for (semantic search)
<k> Number of results (default: 5)

Examples

bash
# Find architecture patterns
agentdb skill search "domain adapter pattern" 5

# Find deployment procedures
agentdb skill search "deploy to raspberry pi" 3

# Find naming conventions
agentdb skill search "naming conventions streams fields" 3

# Find troubleshooting guides
agentdb skill search "mqtt data not appearing" 5

Fallback Method: Reflexion Retrieve

If no skill patterns exist, search past experiences:

bash
agentdb reflexion retrieve "<query>" --k 5 --only-successes --synthesize-context

Parameters

Parameter Description
<query> Task description to find similar work
--k Number of results
--only-successes Only successful episodes
--min-reward Minimum success score (0-1)
--synthesize-context Generate coherent summary

Examples

bash
# Find successful similar work
agentdb reflexion retrieve "HTTP source implementation" \
  --k 5 \
  --only-successes \
  --min-reward 0.7

# Get synthesized context
agentdb reflexion retrieve "timescaledb schema" \
  --k 10 \
  --synthesize-context

Pattern Categories

Category Example Queries
Architecture "domain adapter pattern", "hexagonal architecture"
Data Flow "ingestion pipeline", "bronze silver gold"
Development "add new stream", "implement source trait"
Deployment "docker deployment", "raspberry pi setup"
Troubleshooting "mqtt not working", "parquet write errors"
Conventions "naming conventions", "code organization"

Interpreting Results

Results from skill search include:

Field Meaning
Name Pattern identifier
Description The pattern content
Success Rate How often this pattern succeeded (0-100%)
Uses Number of times used

High-value patterns: Success Rate > 80% AND Uses > 3

Typical Workflow

bash
# 1. Search for existing patterns
agentdb skill search "what I'm about to implement" 5

# 2. If found: Follow the pattern
# 3. If not found: Check reflexion for past experiences
agentdb reflexion retrieve "similar task" --k 5 --only-successes

# 4. After work: Record feedback
agentdb reflexion store "feature-id" "task" 0.9 true "Pattern worked well"

# 5. If you discovered something new: Save it
agentdb skill create "pattern-name" "description" "optional-details"

CRITICAL: Record Pattern Usage

After using a pattern, always use the reflexion skill to record whether it helped:

bash
# Pattern worked well
agentdb reflexion store "dp-004" \
  "Used domain-adapter pattern for new HTTP source" \
  1.0 true \
  "Pattern was complete - followed steps exactly, tests passed"

# Pattern needed fixes
agentdb reflexion store "dp-004" \
  "Used add-stream pattern but needed adjustment" \
  0.6 true \
  "Pattern missing retention field - should update via save-pattern"

Without feedback, the system can't learn which patterns work.

If No Patterns Found

  1. Check pattern stats:

    bash
    agentdb db stats

  2. Search reflexion episodes:

    bash
    agentdb reflexion retrieve "your query" --k 10 --synthesize-context

  3. Check file-based documentation:

    • docs/architecture/ - Architecture documents
    • docs/procedures/ - Step-by-step procedures
    • product/features/*/architecture/ - Feature ADRs

  4. After implementing, store the new pattern via save-pattern

The Pattern Workflow

1. BEFORE work:  get-pattern  → Search for relevant patterns (THIS SKILL)
2. DURING work:  Apply the pattern, note what works/gaps
3. AFTER work:   reflexion    → Record if pattern helped (required)
                 save-pattern → Store NEW discoveries (if any)
                 learner      → Auto-discover patterns from episodes (periodic)

Related Skills

  • save-pattern - Store NEW patterns after discovering reusable approaches
  • reflexion - Record feedback on pattern effectiveness (REQUIRED after using patterns)
  • learner - Auto-discover patterns from successful episodes (user-invoked)

What NOT to Use This For

Don't Search For Use Instead
Current swarm status claude-flow swarm tools
Agent task state claude-flow task tools
Working memory claude-flow memory tools
Session context claude-flow memory with TTL

Patterns are PERMANENT application knowledge, not transient swarm state.

Didn't find tool you were looking for?

Be as detailed as possible for better results