Mirror Palace
A cognitive framework toolkit that teaches AI agents how you work — so they can help you work better.
> Security-first public-repo rule: This repository contains only generalized framework content. Personal/user-specific data must remain segregated in private systems and never be committed here.
---
Why This Exists
AI agents are getting good at remembering what you're working on. Your projects, your tasks, your calendar, your code. That's table stakes.
What they don't remember is how you work. How you make decisions. Where you get stuck. What tends to trigger avoidance. How your stress response affects execution. How your energy moves through the day. Which environments help and which ones create drag.
That second kind of memory, the memory of your patterns, is where the real leverage is.
> Project memory helps you finish things. > Self-knowledge memory helps you build a life that works.
A calendar briefing tells you what's happening today. Mirror Palace gives an agent a structured way to reason about the patterns behind the day.
The difference between a generic briefing and a genuinely useful one is not more raw data. It's better interpretation. If a new project spree usually signals drift, that matters. If low sleep reliably changes what kind of work is realistic, that matters. If a repeated delay pattern shows up around certain decisions, that matters too.
Mirror Palace is the structured knowledge base for that layer. Not a personality quiz. Not a chatbot memory dump. A framework library, status system, and agent toolkit organized so AI systems can support clearer thinking, better decisions, and more grounded action.
---
Architecture
graph TB
classDef knowledge fill:#2d3748,stroke:#4a5568,color:#e2e8f0
classDef status fill:#1a365d,stroke:#2a4365,color:#bee3f8
classDef agent fill:#22543d,stroke:#276749,color:#c6f6d5
classDef action fill:#744210,stroke:#975a16,color:#fefcbf
classDef process fill:#553c9a,stroke:#6b46c1,color:#e9d8fd
subgraph KNOW ["KNOWLEDGE LAYER — What the agent knows about humans"]
F[48 Frameworks] --> EP[Epistemology<br/><i>How to think</i>]
F --> DM[Decision-Making<br/><i>How to decide</i>]
F --> BP[Behavioral Psychology<br/><i>How people act</i>]
F --> CT[Cognitive Therapy<br/><i>Reframing patterns</i>]
F --> EF[Executive Function<br/><i>Focus & initiation</i>]
F --> SI[Self-Image & Identity<br/><i>Who you believe you are</i>]
F --> TR[Trauma Recovery<br/><i>Healing old patterns</i>]
F --> ID[Influence Defense<br/><i>What to watch for</i>]
F --> PA[Personality Assessments<br/><i>Big Five, Enneagram, MBTI</i>]
F --> MORE[+ Coaching, Somatic,<br/>Learning, Anti-Patterns]
end
subgraph DATA ["STATUS LAYER — What the agent knows about you"]
S[Your Life Data] --> DOM[10 Life Domains<br/><i>RYG status + scores</i>]
S --> PPL[People Records<br/><i>Support% · Challenge%</i>]
S --> DEC[Decisions Ledger<br/><i>Pending · Decided · Revisiting</i>]
end
subgraph AGENTS ["AGENT LAYER — Who's watching"]
A[6 Archetypes] --> MR[The Mirror<br/><i>Pattern detection</i>]
A --> BR[The Briefer<br/><i>Daily priorities</i>]
A --> TK[The Tracker<br/><i>Intent vs reality</i>]
A --> WA[The Watcher<br/><i>Change monitoring</i>]
A --> ST[The Strategist<br/><i>Goal pipelines</i>]
A --> OP[The Operator<br/><i>Domain ops</i>]
end
subgraph SKILLS ["ACTION LAYER — How it learns and integrates"]
SK[3 Skills] --> SC[Scan<br/><i>Analyze & populate</i>]
SK --> SE[Setup<br/><i>Integrate with your repo</i>]
SK --> CF[Create Framework<br/><i>Scaffold new frameworks</i>]
end
subgraph DAILY ["PROCESS LAYER — Daily rhythm"]
D[Templates] --> BRF[Morning Briefing]
D --> REF[Reflection]
D --> CHK[Check-in]
D --> FOD[Framework of the Day]
end
IDX[index.md<br/><b>What framework when?<br/>What to update?</b>] -->|routes| F
IDX -->|refreshes| S
AGENTS -->|applies| KNOW
AGENTS -->|monitors| DATA
SC -->|populates| DATA
SC -->|recommends| KNOW
SE -->|connects to| EXT[Your Repo<br/><i>Claude Code · OpenClaw</i>]
class EP,DM,BP,CT,EF,SI,TR,ID,PA,MORE,F knowledge
class DOM,PPL,DEC,S status
class MR,BR,TK,WA,ST,OP,A agent
class SC,SE,SK action
class BRF,REF,CHK,FOD,D process---
How It Works
graph LR
subgraph "1. SCAN"
A[Your docs, messages,<br/>calendar, conversations] --> B[Signal detection]
B --> C[Framework matching]
end
subgraph "2. POPULATE"
C --> D[Status domains updated]
C --> E[People records created]
C --> F[Personality profiles started]
C --> G[Decisions logged]
end
subgraph "3. APPLY"
D & E & F & G --> H[Agent briefings<br/>informed by frameworks]
H --> I[Pattern alerts]
H --> J[Decision support]
H --> K[Relationship insights]
end
subgraph "4. LEARN"
I & J & K --> L[Every interaction<br/>updates the model]
L -->|continuous| A
endThe system gets more useful over time because patterns can be revisited, refined, and connected across decisions, domains, and routines. The goal is not passive accumulation. The goal is better judgment.
The Full Picture
The cycle above shows the mechanics. The diagram below shows how the system thinks — what it detects, how it responds, and what it's ultimately building toward.
graph TB
classDef input fill:#1a365d,stroke:#2a4365,color:#bee3f8
classDef detect fill:#2d3748,stroke:#4a5568,color:#e2e8f0
classDef state fill:#553c9a,stroke:#6b46c1,color:#e9d8fd
classDef respond fill:#22543d,stroke:#276749,color:#c6f6d5
classDef store fill:#744210,stroke:#975a16,color:#fefcbf
classDef vision fill:#742a2a,stroke:#9b2c2c,color:#fed7d7
subgraph INPUT ["WHAT COMES IN"]
I1[Conversation]
I2[Documents]
I3[Email · Calendar · Messages]
I4[Repo activity]
end
subgraph DETECT ["WHAT THE SYSTEM SEES"]
direction TB
D1["Problems & Gaps<br/><i>Avoidance, stalls, drift,<br/>broken streaks, overcommitment</i>"]
D2["Thriving & Flourishing<br/><i>Sustained satisfaction, upward trends,<br/>values-action alignment, creative output</i>"]
D3["Relational Dynamics<br/><i>Support given & received,<br/>reciprocity balance, stakeholder impact</i>"]
D4["Meaning Signals<br/><i>Purpose alignment, legacy potential,<br/>generative work, contribution</i>"]
end
subgraph STATE ["HOW IT READS YOU"]
S1{"User State Check"}
S2["Calm · Receptive"]
S3["Activated · Anxious"]
S4["Flat · Withdrawn"]
S5["Energized · Productive"]
S1 --> S2 & S3 & S4 & S5
end
subgraph RESPOND ["HOW AGENTS RESPOND"]
R1["The Mirror<br/><i>Full reflection: patterns,<br/>meaning questions,<br/>relational awareness</i>"]
R2["The Briefer<br/><i>State-paced briefing:<br/>what's working + priorities +<br/>people to show up for</i>"]
R3["The Tracker<br/><i>Positive trends & streaks<br/>alongside gaps.<br/>Reciprocity tracking.</i>"]
R4["The Strategist<br/><i>Sufficiency checks.<br/>Stakeholder awareness.<br/>Release recommendations.</i>"]
end
subgraph STORE ["WHAT GETS CAPTURED"]
ST1["Domains<br/><i>Issues + Thriving notes</i>"]
ST2["People<br/><i>Support% · Challenge%<br/>Giving% · What I Give</i>"]
ST3["Decisions<br/><i>Who's Affected<br/>Released as valid outcome</i>"]
end
subgraph VISION ["WHAT IT'S ALL BUILDING TOWARD"]
V["Surface patterns → Calm the mind → Act with integrity → Live fully"]
end
I1 & I2 & I3 & I4 --> D1 & D2 & D3 & D4
D1 & D2 & D3 & D4 --> S1
S2 -->|"full depth"| R1
S3 -->|"warmth first,<br/>one insight max"| R2
S4 -->|"gentle,<br/>don't demand"| R2
S5 -->|"match energy,<br/>go deeper"| R1
R1 & R2 & R3 & R4 --> ST1 & ST2 & ST3
ST1 & ST2 & ST3 -->|"continuous learning"| DETECT
R1 & R2 & R3 & R4 --> V
class I1,I2,I3,I4 input
class D1,D2,D3,D4 detect
class S1,S2,S3,S4,S5 state
class R1,R2,R3,R4 respond
class ST1,ST2,ST3 store
class V visionWhat changed from a typical self-knowledge system:
| Traditional Approach | Mirror Palace | |---------------------|---------------| | Detect problems | Detect problems and thriving | | Track what others do for you | Track what you give and receive | | Always analyze | Read your state — back off when you need calm | | Optimize endlessly | Know when enough is enough | | Self-focused | Stakeholder-aware — who else does this affect? | | Clarity as the goal | Clarity in service of a fulfilling life |
---
The 10 Life Domains
Every area of your life gets tracked, scored, and cross-linked.
┌─────────────────────────────────────────────────────────────┐
│ ISSUE INDEX │
│ Cross-domain view · RYG status · Scores · Linked issues │
├──────────┬──────────┬──────────┬──────────┬─────────────────┤
│ FIN-001 │ CAR-003 │ HLT-002 │ PAR-001 │ GRO-004 │
│ 🔴 35 │ 🟡 62 │ 🟢 78 │ 🟡 55 │ 🟢 82 │
└──────────┴──────────┴──────────┴──────────┴─────────────────┘| Domain | Prefix | Tracks | |--------|--------|--------| | Money & Finances | FIN | Burn rate, runway, income streams, targets | | Career & Work | CAR | Role type, stage, leverage score | | Health & Fitness | HLT | Sleep/exercise/nutrition, frequency, energy impact | | Fun & Recreation | FUN | Activity type, social/solo, recharge score | | Environment | ENV | Home/work/third-place, friction level | | Community | COM | Group/scene, role, belonging score | | Family & Friends | FAM | Contact frequency, reciprocity score | | Partner & Love | PAR | Communication quality, polarity, growth trajectory | | Personal Growth | GRO | Domain, format, active/paused, integration level | | Spirituality | SPR | Practice type, frequency, depth, daily integration |
Plus: People Records (support%, challenge%, giving% — what you receive AND what you give) and a Decisions Ledger (status, reversibility, who's affected, domain links, regret check).
---
The Framework Library
48 frameworks across 15 categories. Each one has:
| File | Purpose | |------|---------| | theory.md | Deep explanation of the concept — real depth, not summaries | | template.md | Ready-to-fill worksheet — the framework made personal | | agent-prompt.md | Copy-paste snippet for agents to apply it | | README.md | YAML metadata: when to use, when to avoid, what to update |
Categories at a Glance
EPISTEMOLOGY DECISION-MAKING
├── Concept Formation ├── Reversibility Classification
├── MIRROR Architecture ├── Regret Minimization
└── Information Compression ├── North Star Test
└── Ikigai Diagnostic
BEHAVIORAL PSYCHOLOGY COGNITIVE THERAPY
├── Jobs to Be Done ├── Distortion Detection
├── Behavior Equation ├── Linguistic Reframing
├── Habit Loop Design └── Awareness as Intervention
├── Variable Reward Schedules
├── Loss Aversion EXECUTIVE FUNCTION
└── Identity Reinforcement ├── Executive Function Model
├── ADHD Design Rules
SELF-IMAGE └── Time Blindness
├── Self-Image Cybernetics
├── Teleological Psychology CONTINUOUS LEARNING
└── Systems Over Goals └── Closed-Loop Learning
TRAUMA RECOVERY COACHING
├── Four-F Survival Types ├── Structured Self-Coaching
├── Childhood Emotional Neglect ├── Stories vs Facts
├── Emotionally Immature Parents└── Developmental Stages
├── Family Systems
└── Five Wounds INFLUENCE DEFENSE ⚠️
├── Behavioral Signal Reading
SOMATIC ├── Leverage Point Awareness
├── Subconscious Repatterning └── Manipulation Watchouts
└── Embodied Awareness
PERSONALITY ASSESSMENTS
PATTERN DETECTION ├── Big Five (OCEAN)
├── Failure Modes (7 types) ├── Enneagram (9 types)
└── Psychological Contracts └── MBTI (16 types)
ANTI-PATTERNS INTEGRATED PRACTICE
└── System Anti-Patterns (12) └── Rational Yoga> ⚠️ Influence Defense is explicitly defensive — recognizing techniques used on you, not techniques to use on others.
---
Agent Archetypes
Six pre-built agent personalities. Deploy one or all six.
| Archetype | What It Does | Voice | |-----------|-------------|-------| | The Mirror | Detects patterns across weeks. Names what's working and what you're avoiding. Asks "what is this in service of?" Notices relational imbalances. Paces to your state. | Like a letter from a friend who knows you deeply | | The Briefer | "3 things that matter today." Leads with what's working. Adjusts briefing intensity to your energy and emotional state. | Sharp chief of staff | | The Tracker | "You said X. You did Y." Reports positive trends and sustained streaks alongside gaps. Tracks relational reciprocity. | Terse, factual, no judgment | | The Watcher | Monitors changes across repos, docs, status. Flags drift. | Methodical completist | | The Strategist | Tracks goal pipelines. Flags stalls. Recommends releasing goals that no longer serve. Asks "who else does this matter to?" | Strategic but direct | | The Operator | Domain-specific ops. Watches signals, drafts responses. Notes stakeholder impact. | Operational, brief |
Each archetype has a SOUL.md (personality + scope), MEMORY.md (what it knows), HEARTBEAT.md (when it runs), and a README.md explaining when to deploy it.
---
The Scan Skill — How It Learns About You
Four modes, all feeding the same system:
graph TD
D[📄 Documents<br/><i>Journals, notes, plans</i>] --> SCAN
C[💬 Conversation<br/><i>Interactive interview</i>] --> SCAN
M[📡 Connected<br/><i>Email, Calendar, Slack</i>] --> SCAN
R[📂 Repo<br/><i>Existing config files</i>] --> SCAN
SCAN[SCAN ENGINE<br/>Signal → Framework matching] --> OUT
OUT --> S1[Status domains populated]
OUT --> S2[People records created]
OUT --> S3[Personality profiles started]
OUT --> S4[Framework recommendations]
OUT --> S5[Agent archetype suggestion]Continuous Learning Protocol: The scan isn't a one-time setup. Every conversation is a potential data point. The system proposes low-friction updates as you talk — one line at a time, confirm or skip. Over weeks and months, it builds a progressively richer model of who you are and how you operate.
---
Quick Start
| Path | Time | Start Here | |------|------|-----------| | Browse frameworks | 5 min | index.md — the master routing table | | Understand yourself | 15 min | Run the scan skill on your existing documents | | Build an agent | 10 min | agents/archetypes/ — pick one | | Integrate with your repo | 10 min | guides/getting-started.md | | Daily practice | 3 min/day | daily/briefing-template.md |
---
Integration
Works with Claude Code and OpenClaw.
┌─────────────────────────────────┐
│ YOUR REPO │
│ AGENTS.md ← references MP │
│ CLAUDE.md ← continuous learn │
│ agents/ ← deployed types │
│ memory/ ← agent memory │
├─────────────────────────────────┤
│ ↕ scan populates ↕ setup │
├─────────────────────────────────┤
│ MIRROR PALACE │
│ frameworks/ (knowledge) │
│ status/ (your data) │
│ agents/ (archetypes) │
│ skills/ (scan+setup+create) │
│ daily/ (process) │
│ index.md (routing table) │
└─────────────────────────────────┘See guides/claude-code-setup.md or guides/openclaw-setup.md.
---
Philosophy
The system optimizes for clarity in service of flourishing — not just seeing yourself clearly, but using that clarity to build a life worth living.
Surface patterns → Calm the mind → Act with integrity → Live fullyEvery component is oriented toward this progression: agents detect what's working alongside what's broken, adjust their intensity to your state, track relational reciprocity (not just what you receive), consider who else your decisions affect, and know when enough is enough.
See docs/ROADMAP.md for the full philosophy and how it's applied.
---
The Deeper Point
Most productivity tools optimize for output: more tasks done, more habits tracked, more data collected.
Mirror Palace optimizes for understanding. The bet is simple: an agent that deeply understands how you operate — your patterns, your wounds, your strengths, your decision-making tendencies, your relationship dynamics, your energy rhythms — can do more for you in a 3-minute morning briefing than a feature-rich task manager can do in a year.
This isn't about building a better to-do list. It's about building a system that knows:
- When you say "I'll start Monday" you mean "I'm avoiding this"
- When your calendar fills up it's not ambition, it's over-commitment
- When you go quiet in relationships it's a pattern, not a preference
- When you compare yourself to others it means your self-worth is untethered from your actual work
- When you start something new with intense energy, the follow-through is where the real test is
That knowledge — structured, accessible, continuously updated — is what turns a helpful assistant into a genuine thinking partner. One that doesn't just tell you what's on your calendar, but helps you build a life with less friction, less self-deception, and more of the clarity that actually moves things forward.
---
Acknowledgments
This project grew out of too many late nights, too many frameworks studied, and too many projects running at once. It draws from the memory palace tradition, graph-based knowledge systems like Graphify, and every psychology, behavioral science, coaching, and self-development framework I've ever studied and actually applied — there are genuinely too many to list.
If you recognize your work in here and feel it needs a credit, reach out. The frameworks are described by concept, not by name, because this is a toolkit for application, not an academic citation index. But the ideas belong to the people who developed them, and this project wouldn't exist without their work.
---
License
MIT
