Memory Management Skill

Use AutoMem to maintain persistent context across Claude Code sessions.

Tool examples below use short names (recall_memory, store_memory); call them on whichever AutoMem MCP server is wired in (plugin installs namespace them as mcp__plugin_automem_memory__*, user-level servers as mcp__memory__*).

This skill teaches the current AutoMem playbook: Recall early, store durable outcomes, avoid session-summary noise.

Phase 1: Session Start (Recall)

Always Recall For

• Project context questions (architecture, tooling, deployment)
• Architecture discussions or decisions
• User preferences and code style
• Debugging issues (search for similar past problems)
• Refactoring (understand why current structure exists)
• Integration or API work (check past implementations)
• Performance optimization discussions

Standard Recall Pattern

• Preferences first: recall with tags: ["preference"], limit: 20, sort: "updated_desc"
• Task context second: one semantic query built from actual nouns in the request, optional project slug if unambiguous
• Debugging on demand: recall with the error text as the semantic query and NO tags — bugfix/solution tagging is incomplete, and a tag gate hides cross-corpus fixes

Skip Recall For

• Pure syntax questions ("How does Array.map work?")
• Trivial edits (typos, formatting, simple renames)
• Direct factual queries about current code
• File content requests that can be answered by reading

Recall Examples

// Preferences first
recall_memory({
  tags: ["preference"],
  limit: 20,
  sort: "updated_desc"
})

// Task context recall
recall_memory({
  query: "authentication timeout PostgreSQL auth.ts retry logic",
  tags: ["myapp"],   // drop if ambiguous
  time_query: "last 90 days",
  limit: 30
})

// Debug similar errors (no tag gate — a hard gate hides cross-corpus fixes)
recall_memory({
  query: "TimeoutError authentication request timed out",
  limit: 20
})

Phase 2: During Work (Store)

What to Store with Importance Levels

Storage Format

Content: "Brief title. Context and details. Impact/outcome."
Tags: [category, project-slug, language]
Type: Decision | Pattern | Insight | Preference | Style | Habit | Context

Use bare tags only. Do not add platform tags or date tags.

Storage Examples

Decision:

store_memory({
  content: "Chose PostgreSQL over MongoDB. Need ACID guarantees for transactions. Impact: Ensures data consistency.",
  type: "Decision",
  tags: ["decision", "myapp", "database"],
  importance: 0.9,
  confidence: 0.9,
  metadata: {
    alternatives_considered: ["MongoDB", "DynamoDB"],
    deciding_factors: ["ACID", "relationships", "team_expertise"]
  }
})

Bug Fix:

store_memory({
  content: "Auth timeout on slow connections. Root: Missing retry logic. Solution: Added exponential backoff with 3 retries.",
  type: "Insight",
  tags: ["bugfix", "solution", "myapp", "auth"],
  importance: 0.8,
  confidence: 0.85,
  metadata: {
    error_signature: "TimeoutError: Authentication request timed out",
    solution_pattern: "exponential-backoff-retry",
    files_modified: ["src/auth/client.ts"]
  }
})

User Preference:

store_memory({
  content: "User prefers early returns over nested conditionals in validation code.",
  type: "Preference",
  tags: ["preference", "code-style"],
  importance: 0.8,
  confidence: 0.95
})

After Storing: Create Associations

Link related memories to build a knowledge graph:

associate_memories({
  memory1_id: "related-memory-id",
  memory2_id: "new-memory-id",
  type: "INVALIDATED_BY",  // or PREFERS_OVER, LEADS_TO, EXEMPLIFIES
  strength: 0.9
})

Prefer update_memory over near-duplicate stores when a fact changed in place.

Best Practices

Do

• Load context automatically at session start
• Store high-signal events (decisions, bugs, patterns)
• Create specific relationship types when the link is explicit
• Include rich metadata in memories
• Present recalled information naturally
• Tag consistently with bare tags only

Don't

• Store secrets, API keys, or sensitive data
• Store trivial changes (typos, formatting)
• Create associations without verifying relevance
• Use platform tags or date tags
• Announce "I'm searching my memory" constantly
• Store large code blocks (store patterns/decisions instead)
• Store end-of-session summaries by default

Natural Integration

When recalling memories, weave context seamlessly into responses. Avoid robotic phrases like "searching my memory database" and present memories as normal working context.