Skill in 《Session Dream》

Skill Description

On-demand session memory distillation — extracts key decisions, eliminated approaches, new discoveries, and current blockers from the current conversation and writes them to MEMORY.md topic files. The manual trigger for Claude Code's autoDream background consolidation service.

Skill.md

Session Dream

Claude Code ships with an autoDream background service that fires after 5+ sessions spanning 24+ hours. It scans historical conversation logs and consolidates insights into your project's MEMORY.md topic files automatically. Useful — but not immediate.

If you spend three hours debugging a subtle issue and resolve it today, autoDream won't capture it until the threshold is met. If context compaction happens tonight, the reasoning behind what you learned disappears into the summary. Next session, Claude starts fresh with no memory of what you worked through.

session-dream closes that gap: trigger memory distillation on demand, without waiting for autoDream to run.

Four-Phase Distillation Process

PhaseWhat happens
OrientRead existing MEMORY.md — map what's already covered, avoid duplicates
GatherScan the current conversation — classify signals by value tier
ConsolidateWrite or update topic files — refine contradictions rather than overwrite
Prune & IndexUpdate MEMORY.md pointers — keep it under the 200-line limit

Signal Value Tiers

High-value — always distill:

  • Architectural decisions with reasoning — "chose X over Y because Z"
  • Dead ends — "tried A, hit wall B, don't retry" is often more valuable than the solution
  • Corrections to prior memory
  • Constraints not written anywhere — implicit rules that only exist in someone's head

Medium-value — distill if not easily discoverable:

  • Project-specific naming conventions
  • Environment-specific quirks
  • Current blockers and their root cause

Always skip:

  • File paths read, commands run
  • Implementation details visible in the code
  • Step-by-step narration of what was done

How to Trigger

  • Say dream or /dream
  • Say "save session memories", "distill this session", "what should I remember from this session?"
  • At the end of any long, productive session before switching context

Relationship to Other Memory Skills

session-dream is on-demand distillation: run it when a session wraps up. compact-with-memory is pre-compression distillation: run it when the context window is filling up. memory-architect is structural cleanup: run it when MEMORY.md has accumulated redundant or stale entries.

The three skills cover different points in the memory lifecycle and work best together.

Limitation

Distillation quality depends on how explicitly decisions were articulated during the session. If reasoning wasn't stated — just implied — the memory entries will lack the **Why:** context that makes them useful. A practical habit: when making an important architectural call, say the reason out loud so it can be captured when you dream.

Install & Use

Install command

npx skills add simbajigege/book2skills/skills/session-dream
OR

Direct download

Related Skills

Tool Permission System AI Skill

Design a configurable, layered permission/safety system for agent tools: one permission pipeline decides allow / ask / deny across enterprise/user/project scopes.

Agent Memory Implementation Guide AI Skill

Restructures a bloated or chaotic MEMORY.md into the 2-layer architecture Claude Code's autoDream service uses intern...

Compact with Memory AI Skill

Executes /compact correctly — generates a high-quality conversation summary that preserves reasoning, decisions, and ...

Harness Step 1 — Create AGENTS.md & Knowledge Base AI Skill

Scans your codebase and generates AGENTS.md (the agent's orientation file) plus a complete docs/ knowledge base skele...

Compact Memory Implementation Guide AI Skill

A developer implementation guide for adding compact memory to an Agent built with the Anthropic API or Claude Agent S...

Harness Step 3 — Session State Management AI Skill

Creates init.sh (environment health check), tasks.json (structured task queue), and progress.md (append-only session ...

LangChain Tool Builder AI Skill

Build LangChain (Python) tools using Claude Code's fail-closed design pattern — unified name, schema, security, and e...

Harness Step 2 — Fill docs/ with Real Content AI Skill

Deep-reads your codebase and fills every docs/ file with specific, accurate content grounded in what the code actuall...

Query Loop Implementation AI Skill

Implement a production-ready LLM query loop for AI applications: tool calling, structured tool_result feedback, ReAct...

Creating Superfans AI Skill

Apply Brittany Hodak's SUPER Model from Creating Superfans to turn ordinary customers into advocates through story, c...

Essays in Persuasion: Keynes Policy Reasoning AI Skill

Apply Keynes's Essays in Persuasion to evaluate macro policy, debt capacity, inflation-deflation tradeoffs, fixed-rat...