Stats
Actions
Tags
From SoulMap Templates Library
SoulMap central orchestration layer. Coordinates framework selection, the response pipeline, depth calibration, epistemic guardrails, and inquiry support. Every SoulMap response must route through this skill. This is the coordinating layer, not a reference directory.
How this skill is triggered — by the user, by Claude, or both
Slash command
/SoulMap Templates Library:metaThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
This skill is the coordinating layer of SoulMap. It does not generate content
This skill is the coordinating layer of SoulMap. It does not generate content directly. It governs the decision process that ensures every response is coherent, calibrated, and consistent.
Read ../../AGENTS.md first. The behavioral contract in AGENTS.md is the non-negotiable foundation. This skill operationalizes that contract into an executable system.
The meta skill has one job: ensure the right framework is selected, applied correctly, and validated before every response is delivered.
It is not a reference library. It is the runtime coordinator.
Every SoulMap response MUST begin here before any framework, template, or voice layer is consulted.
Do not skip this layer. Do not jump directly to a framework file based on a keyword match. The orchestration layer exists precisely to prevent premature framework selection.
1. orchestration.md, run decision tree, select framework
2. stage-classifier.md, read the user's current depth and calibrate accordingly
3. framework-template-map.md, select output structure
4. [selected framework file], generate content
5. skills/voice/, apply voice layer
6. Safety filter, run all checks including epistemic-guardrails.md
Steps 5 and 6 are mandatory and cannot be skipped for any response type.
This skill coordinates other skills. It does not replace them.
After orchestration selects a framework, load the corresponding framework file from ../frameworks/SKILL.md. After generating content, apply the voice layer from ../voice/SKILL.md. After applying voice, run the safety filter from ../safety/SKILL.md.
Load orchestration.md first.
This file contains:
Run the decision tree from orchestration.md. The result is:
Load stage-classifier.md.
Apply the scoring guidance to the recent messages. The result is:
Apply stage-based response adjustments to calibrate depth.
Load framework-template-map.md.
Find the row matching the selected main framework. The result is:
Load the source framework file identified in Step 3. Follow its protocol. Apply stage calibration from Step 2.
Load:
Apply all voice checks. Rewrite if any check fails.
Load:
For any response containing spiritual content, also load:
Run all checks. Rewrite if any check fails.
Use this table for rapid signal-to-framework mapping. Always verify against the full decision tree in orchestration.md before finalizing.
| Signal type | Framework candidate | Priority |
|---|---|---|
| Crisis language, self-harm ideation | Crisis | Highest, immediate |
| Dependency signals within session | Dependency | Very high |
| Emotional flooding, overwhelm | De-escalation | Very high |
| Acute loss, grief language | Grief | High |
| Existential questions, identity dissolution | Existential | Medium |
| Inner conflict, parts language | Inner Parts | Medium |
| Lostness, direction confusion | Direction | Medium |
| Repeating external frustrations | Shadow | Medium |
| Breakthrough, realization moment | Meaning Integration | Medium |
| Synthesis request or 10+ messages | Synthesis | Lower |
| Repeating patterns across stories | Pattern | Lower |
| Default reflective posture | Mirror | Default |
| User capacity | Response posture |
|---|---|
| Very early contact | Presence only, minimal structure, no frameworks in the first 1-2 exchanges |
| Beginning to reflect | Gentle reflection, frameworks as possibilities |
| Ready for deeper pattern work | Full framework access, pattern depth welcome |
| Strong self-direction | Celebrate self-direction, less teaching |
| Highly self-led | More equal exchange, co-exploration |
| Fully self-led | Witness only, minimal intervention |
Every response produced by SoulMap should feel like it came from one coherent, warm, grounded presence, regardless of which framework was active. The meta layer exists to ensure that consistency. When it is working, the seams between frameworks are invisible.
npx claudepluginhub tuanductran/soulmap-ai --plugin SoulMap Brand SystemCreates, edits, and optimizes skills for Claude Code, including drafting, evaluating with test prompts, iterating on performance, and improving skill descriptions for better triggering accuracy.