Resources
10Install
npx skillscat add 1306885728/self-mastery-skill Install via the SkillsCat registry.
SKILL.md
/{{SLUG}}
Open-source, single-user, local-first self-mastery skill.
This template is meant to be forked. Replace {{SLUG}} and {{YOUR_NAME}}, then follow SETUP.md.
Routing
| Trigger | File | Purpose |
|---|---|---|
/{{SLUG}} |
SKILL.md |
Default entry point |
/{{SLUG}}-business |
modes/business.md |
Business and decision support |
/{{SLUG}}-decide |
modes/business.md |
Alias for decision support |
/{{SLUG}}-emotion |
modes/emotion.md |
Emotional support |
/{{SLUG}}-reach |
modes/reach.md |
External research / reach mode |
/{{SLUG}}-journal |
modes/journal.md |
Read-only journal review |
/{{SLUG}}-memory |
modes/memory.md |
Memory search |
/{{SLUG}}-update |
modes/update.md |
Confirmed archive updates |
/{{SLUG}}-search |
modes/search.md |
Internal full-text search |
/{{SLUG}}-self |
modes/self_mode.md |
Self archive view |
/{{SLUG}}-persona |
modes/persona_mode.md |
Persona rules view and tuning |
/{{SLUG}}-import |
modes/import.md |
Safe import wizard |
Core Rules
- Treat this skill as a diary and insight system, not a role-play bot.
- Keep
self.mdandpersona.mdfor human-confirmed long-term archive content only. - Write automatic ingestion and analysis into the local diary layer first:
{skill_root}/diary/{skill_root}/growth/reports/{skill_root}/growth/state/- The actual runtime paths are created from local config on first run.
- Only write stable conclusions back into
self.mdorpersona.mdafter explicit confirmation or correction. - Keep analysis local by default. Cloud enhancement must stay opt-in and source-limited.
Working Style
- For imports, records, reviews, and summaries, prefer local files and the local API.
- For "how have I been lately" questions, read daily snapshots, trend windows, and weekly reports instead of guessing.
- For correction requests, record a correction first and let future analysis reference it.
- Do not present automatic analysis as confirmed archive truth.
- If emotional keywords appear, switch to
modes/emotion.mdautomatically.
Automatic Switching
| Detected content | Auto mode |
|---|---|
| Emotional keywords such as lonely, upset, overwhelmed | modes/emotion.md |
| Clear business topics such as partnership, funding, contract | modes/business.md |
| Clear correction such as "that is wrong" or "you misread me" | modes/update.md correction flow |
If nothing clearly matches, stay in the default conversation mode.
Growth API
Check the local API before relying on growth-layer data:
GET http://127.0.0.1:8765/health
Authorization: Bearer <local token>If the API is unavailable, fall back to local diary files. If those are also unavailable, tell the user to start the daemon.
Privacy and Cloud
- Default mode is
analysis.mode=local. privacy.cloud_enabled=falsemeans no raw data should be sent to external models.- Cloud enhancement, if enabled, should only receive allowed sources or summarized, de-identified content.
- Every external send should be written to a local audit log.
Quick Start
python -m growth.daemon
python -m growth.daemon --onceFirst run creates the local config and local output directories automatically.