From dream-mentor
**Description**: Core dream state lifecycle management - load, analyze, store, check, update, and diff reference codebases.
npx claudepluginhub issacchaos/local-marketplace --plugin dream-mentorThis skill uses the workspace's default tool permissions.
**Description**: Core dream state lifecycle management - load, analyze, store, check, update, and diff reference codebases.
Searches, retrieves, and installs Agent Skills from prompts.chat registry using MCP tools like search_skills and get_skill. Activates for finding skills, browsing catalogs, or extending Claude.
Searches prompts.chat for AI prompt templates by keyword or category, retrieves by ID with variable handling, and improves prompts via AI. Use for discovering or enhancing prompts.
Guides MCP server integration in Claude Code plugins via .mcp.json or plugin.json configs for stdio, SSE, HTTP types, enabling external services as tools.
Description: Core dream state lifecycle management - load, analyze, store, check, update, and diff reference codebases.
Trigger: Used by the /mentor command for all dream operations.
The dream-manager skill orchestrates the full lifecycle of dream states:
.claude/dreams/<name>/| Sub-Skill | File | Purpose |
|---|---|---|
| Loader | loader.md | Source resolution, cloning, path validation |
| Analyzer | analyzer.md | Parallel repo analysis (structure, patterns, deps, conventions) |
| Memory Writer | memory-writer.md | Write/update dream memory files and index |
| Staleness Checker | staleness-checker.md | Detect when source repo has new commits |
| Diff Engine | diff-engine.md | Compare current project against dream state |
Input: source (URL or path), name, layer priority, branch
│
├─ loader.md: Resolve source type, validate access
│ ├─ GitHub: gh repo clone → temp dir
│ └─ Local: validate path exists, is git repo
│
├─ analyzer.md: Run parallel analysis agents
│ ├─ Agent 1: Structure analysis (dirs, modules, entry points)
│ ├─ Agent 2: Pattern analysis (architecture, naming, code style)
│ ├─ Agent 3: Dependency analysis (package files, frameworks)
│ └─ Agent 4: Convention analysis (testing, build, CI, docs)
│
├─ memory-writer.md: Write memory files
│ ├─ source.json (metadata)
│ ├─ layer.json (priority config)
│ ├─ structure.md
│ ├─ patterns.md
│ ├─ dependencies.md
│ ├─ conventions.md
│ └─ summary.md
│
└─ memory-writer.md: Update DREAMS.md index
Output: Dream loaded confirmation with summary
Input: dream name (or all dreams)
│
├─ Read source.json for each dream
├─ staleness-checker.md: Compare last_commit vs current HEAD
│ ├─ Local: git -C <path> rev-parse HEAD
│ └─ GitHub: gh api repos/<owner>/<repo>/commits/<branch>
│
└─ Return: { stale: bool, commits_behind: int, changed_files: [] }
Input: dream name (or all stale dreams)
│
├─ staleness-checker.md: Get changed files since last_commit
│ ├─ Local: git -C <path> diff --name-only <last_commit>..HEAD
│ └─ GitHub: gh api compare endpoint
│
├─ analyzer.md: Re-analyze only affected areas
│ ├─ Map changed files → memory categories
│ └─ Re-run relevant analysis agents
│
├─ memory-writer.md: Update affected memory files
│
└─ memory-writer.md: Update source.json (new commit, timestamp)
Input: dream name (or merged layer view)
│
├─ Read dream memory files
├─ diff-engine.md: Analyze current project
├─ diff-engine.md: Compare dream vs current
│ ├─ Structure differences
│ ├─ Pattern divergences
│ ├─ Dependency gaps
│ └─ Convention mismatches
│
└─ Output: Diff report with recommendations
The analyzer produces 5 memory files per dream:
Dreams are stored per-project in .claude/dreams/. No global configuration is needed.
None required. The skill uses gh CLI for GitHub operations (must be authenticated).
git CLI available in PATHgh CLI available and authenticated (for GitHub repos)