memory-extraction

name	memory-extraction
description	Smart observation capture - automatically identify and record important patterns, decisions, and insights during work. Use proactively during development sessions.
user-invocable	false

Memory Extraction

Automatically identify and record important observations during a development session.

When to Extract

Record an observation when you encounter:

Error Patterns

• A command fails with a specific error
• A build breaks due to a dependency issue
• A test fails with an unexpected result
• An API returns an error response

Format: [ERROR] <tool/command>: <error message> → <resolution>

Success Patterns

• A debugging approach that worked
• A configuration that resolved an issue
• A workflow that was particularly efficient

Format: [SUCCESS] <what worked> in context of <what was being done>

User Preferences

• The user explicitly states a preference ("always use...", "never do...")
• The user corrects your approach, indicating a preference
• The user chooses between alternatives

Format: [PREFERENCE] <preference description>

Decisions

• Architectural choices (library selection, pattern choice)
• Configuration decisions (environment setup, tool config)
• Workflow decisions (branching strategy, testing approach)

Format: [DECISION] <decision> because <rationale>

Insights

• Patterns you notice in the codebase
• Relationships between components
• Performance characteristics

Format: [INSIGHT] <observation>

How to Record

Call pluggedin_memory_observe with:

• sessionUuid: Current session UUID
• observationType: One of error_pattern, success_pattern, user_preference, decision, insight, tool_call, tool_result, workflow_step
• content: The observation text
• outcome: "success", "failure", or "neutral"
• metadata: Optional JSON with tool names, error codes, etc.