Maintains network health by validating memory connections and performing automated diagnostic repairs.

Install

mkdir -p .claude/skills/dream-state && curl -L -o skill.zip "https://agentskills.codes/api/skills/download/10485" && unzip -o skill.zip -d .claude/skills/dream-state && rm skill.zip

Installs to .claude/skills/dream-state

Activation

This is the description your AI agent reads to decide when to run this skill — the better it matches your request, the more reliably it fires.

Neural maintenance, synapse validation, automated health diagnostics, and unconscious processing
96 charsno explicit “when” trigger
Intermediate

Key capabilities

  • Execute automated synapse validation across memory files
  • Detect and repair broken connections using consolidation mappings
  • Generate health reports with network statistics
  • Synchronize with Global Knowledge repository
  • Classify network health as healthy or attention required

How it works

The skill performs a 6-phase protocol that scans memory files, validates embedded YAML synapses, repairs broken links via consolidation mappings, and generates a status report.

Inputs & outputs

You give it
Memory files in the architecture
You get back
Markdown health report in .github/episodic/

When to use dream-state

  • Running weekly neural health checks
  • Repairing broken internal memory references
  • Synchronizing memory with global knowledge
  • Generating network health reports

About this skill

Dream State Automation Skill

Expert in automated neural maintenance, synapse validation, health diagnostics, and unconscious cognitive processing.

Capabilities

  • Execute automated synapse validation across all memory files
  • Detect and repair broken connections using consolidation mappings
  • Generate comprehensive health reports with network statistics
  • Synchronize with Global Knowledge repository
  • Provide real-time progress feedback during maintenance
  • Classify network health (HEALTHY vs ATTENTION REQUIRED)

When to Use This Skill

  • User asks to "dream" or run "neural maintenance"
  • After file reorganizations that may break synapse references
  • Regular weekly health checks
  • Before major meditation sessions (clear cognitive overhead)
  • After domain learning to validate new connections
  • When architecture health is uncertain before SSO planning

Core Concepts

6-Phase Dream Protocol

PhaseFunctionOutput
1. DiscoveryScan .github/ for all memory filesFile inventory
2. ValidationParse and verify all embedded synapsesBroken connection list
3. RepairAuto-fix using consolidation mappingsRepair log
4. GK SyncSynchronize with Global Knowledge repoSync status
5. ReportingGenerate health report with statisticsMarkdown report
6. DisplayShow results and open reportVS Code notification

Health Metrics

MetricDescriptionHealthy Threshold
Total Memory FilesAll .md files in architectureTracked, not thresholded
Total SynapsesAll embedded connectionsTracked, not thresholded
Broken ConnectionsSynapses pointing to non-existent files0
Repaired ConnectionsFixed by consolidation mappingsInformational

Consolidation Mappings

Deprecated-to-current file resolution for automatic repair:

  • Renamed files → current names
  • Merged files → unified equivalents
  • Split files → appropriate successors

Dream vs Meditation Distinction

AspectDreamMeditation
ModeUnconscious, automatedConscious, manual
FocusMaintenance, diagnosticsKnowledge consolidation
Creates files?Reports onlyMemory files + synapses
ActivationPrompt or script (brain-qa.cjs)User-requested
AnalogyREM sleepContemplative meditation

Quality Principles

  • Non-destructive: Never deletes memory files
  • Reversible: All changes tracked and reviewable
  • Conservative: Prioritizes system stability
  • Comprehensive: Reports on entire architecture
  • Safe: Creates repair history for audit trails

Example Prompts

  • "Run neural maintenance"
  • "Check architecture health"
  • "Are my synapses healthy?"
  • "Dream — validate connections"

Synapse Format

Synapses are embedded YAML blocks linking memory files:

---
synapses:
  - target: "skills/research-first-development/SKILL.md"
    relation: "applies-to"
    strength: 0.9
  - target: "instructions/debugging-patterns.instructions.md"
    relation: "extends"
    strength: 0.7
---

Valid relations: applies-to, extends, requires, contradicts, supersedes

Dream Report Template

# Dream Report — 2026-04-14

## Summary
- **Files Scanned**: 168
- **Synapses Validated**: 412
- **Broken Connections**: 0
- **Auto-Repaired**: 2
- **Status**: HEALTHY ✓

## Repairs Applied
| File | Broken Target | Repaired To |
|------|---------------|-------------|
| skills/api-design/SKILL.md | old-path.md | new-path.md |

## Recommendations
None — architecture is healthy.

Input Expectations

  • No input required (zero-configuration)
  • Optional: specific file or folder to focus on

Output Format

  • Health report in .github/episodic/dream-report-*.md
  • Summary statistics (files, synapses, broken, repaired)
  • Recommendations for manual fixes if needed
  • Network health status: HEALTHY or ATTENTION REQUIRED

When not to use it

  • When the user specifically requests manual meditation
  • When architecture health is already confirmed as healthy

Limitations

  • Does not delete memory files
  • Requires consolidation mappings for automatic repairs

How it compares

Unlike manual meditation which focuses on knowledge consolidation, this skill automates maintenance and diagnostics without creating new memory files.

Compared to similar skills

dream-state side by side with the closest alternatives in the catalog.

SkillInstallsUpdatedSafetyDifficulty
dream-state (this skill)03moNo flagsIntermediate
yams12moReviewIntermediate
trade04moNo flagsIntermediate
scientific-brainstorming377moNo flagsIntermediate

Try saying

Example prompts that trigger this skill in your AI assistant.

Search skills

Search the agent skills registry