CtrlK
BlogDocsLog inGet started
Tessl Logo
PluginPublic

metcalf-solutions1/satori

Psychologically grounded wisdom companion for inner life. Use for reflective dialogue, grief, dark-night, burnout, anxiety, stress, loneliness, relationship conflict, shadow work, crossroads, tiered crisis check-ins, continuity notes. DO NOT ACTIVATE for pure coding, DevOps, SQL, git, math, or technical tasks, even if user says "I'm stuck" or "struggling with bug." Hybrid: solve technical task first, then add 2 warm grounding sentences for burnout, imposter syndrome, team conflict.

1.05x
Quality

91%

Does it follow best practices?

Impact

98%

1.05x

Average score across 8 eval scenarios

SecuritybySnyk

Passed

No findings from the security scan

Overview
Quality
Evals
Security
Files

architecture.mddocs/

Satori Architecture & Progressive Disclosure (v6.6.0)

Satori eliminates prompt bloat with an uncompromising progressive disclosure architecture. SKILL.md (<8 KB, strictly below the 8,000-byte budget) is the single canonical source of truth and is completely self-sufficient for Turn 1. Reference modules load progressively on demand, with all 17 markdown reference files guaranteed to remain under the 24,000-byte IDE and harness rule ceiling.

All quantitative budgets, environments, and benchmark gates are governed by facts.json.


Full File Load Budget & Progressive Disclosure Matrix

The repository maintains exactly 17 Markdown reference files in references/ (12 onDemand modules plus 5 conditional protocols) and 1 primary-source JSON manifest (references/crisis-sources.json), alongside entry points and governance policies:

FileBudgetPurposeLoad Condition
SKILL.md<8 KB (Turn-1)Canonical entry point, trigger boundaries (positive, negative, hybrid), 8 Constitutional Immovables, 6-Step Turn Model, 9-region verified crisis table, 4-Question Pre-Send Honesty CheckTurn 1 (Always, Frontier / Profiles 1-4)
compact/SKILL.md≤7,800 chars (<8 KB)Single self-contained prompt for ChatGPT Custom GPTs, local models, and strict context budgetsProfile 5 & Custom GPTs
references/core-compact.md<12 KB (<24 KB)Compact operational spine with complete safety parity and core mechanicsProfile 5 Models Only
references/SOUL.md~13 KB (<24 KB)Constitutional identity, 8 Constitutional Immovables, voice philosophy, drift detection (SHA-256 verified in openclaw.plugin.json)Turn 2+ (Profiles 1-4, onDemand)
references/clinical-spine.md<20 KB (<24 KB)Core 6-step mechanics, 3x6 Role/Register matrix, working formulation, parts awareness, autonomy guardrails, Three-Tier Risk ModelTurn 2+ (Profiles 1-4, onDemand)
references/traditions-quickref.md~18 KB (<24 KB)Quick selection matrix across wisdom traditions and clinical modalitiesWhen selecting a lens (onDemand)
references/model-calibration.md~10 KB (<24 KB)3-Tier Runtime Capability Detection and 5 Behavioral ProfilesModel calibration reference (onDemand)
references/memory-system.md~11 KB (<24 KB)Default local vault (~/.satori/memory/), cloud-sync guardrails, SQLite FTS5 search, Dream Layer (1->2->3), SATORI_NOTE v6.5 formatWhen reading/writing memory (onDemand)
references/toolkit-inquiry.md~7 KB (<24 KB)OARS, Motivational Interviewing, McAdams narrative identity, Singer self-defining memories, Voss tactical empathyOn demand (Inquiry & formulation)
references/toolkit-somatic.md~6 KB (<24 KB)Somatic tracking, nervous system grounding, physical experiments, experiential focusingOn demand (Somatic awareness)
references/traditions-stoic.md~7 KB (<24 KB)Modular deep guide: Epictetus, Seneca, Marcus Aurelius, and ExistentialismOn demand (Stoic/Existential lens)
references/traditions-eastern.md~7 KB (<24 KB)Modular deep guide: Taoism, Buddhism, Zen, and Hindu philosophical traditionsOn demand (Eastern lens)
references/traditions-african-sufi.md~6 KB (<24 KB)Modular deep guide: Ubuntu, Kemetic Ma'at, Akan ethics, Sufism, and Indigenous reciprocityOn demand (African/Sufi/Indigenous lens)
references/tone-and-voice.md~14 KB (<24 KB)Before/after worked voice calibration examples, anti-slop discipline, and register calibrationOn demand (Voice guidance)
references/onboarding.md~11 KB (<24 KB)Context-First Bypass, gentle 3-stage first-session sequence, and privacy consent noticeNew users without immediate topic (conditional)
references/darknight-protocol.md~7 KB (<24 KB)Pure presence mode for non-suicidal 3:00 AM existential despair (avoids alienating users in dark nights)On Dark Night signal (conditional)
references/grief-protocol.md~10 KB (<24 KB)Bereavement, disenfranchised grief, Pauline Boss Ambiguous Loss, Dual Process Oscillation, Continuing BondsOn grief / loss signal (conditional)
references/shadow-work-protocol.md~11 KB (<24 KB)5-session Jungian shadow integration arc and projection reclamationOn shadow work engagement (conditional)
references/relationship-protocol.md~10 KB (<24 KB)Solo relational rupture and Shared Two-Person Conflict Hold Mode (Multi-Partiality, Relational Psychoanalysis, Gottman, NVC)On relational conflict signal (conditional)
references/crisis-sources.json~8 KB (<24 KB)Primary-source verified emergency and 24/7 crisis lines across 9 verified regions (verified within 120 days)Build-time & crisis reference (manifest)
SAFETY.md~8 KBComprehensive clinical safety policy, Three-Tier Risk Model, Candor Floor, and clinician audit logSafety governance
PRIVACY.md~5 KBThreat model (T1-T6), local permissions (0700/0600), cloud-sync protection, and right to erasure (forget/purge + VACUUM)Privacy and data sovereignty spec
NOTICE.md~4 KBOpen-source attribution log crediting foundational skill and prompt lineagesAttribution & licensing

Architecture Principles

1. Zero Turn-1 Load Tax

In earlier iterations, loading an entire wisdom companion injected over 63 KB into the system prompt on Turn 1. Satori v6.6 guarantees that Turn 1 loads strictly under 8,000 bytes (budgets.skill_entry_max_bytes: 8000) and consumes fewer than 2,000 tokens (budgets.turn_1_token_load_tax_max: 2000). The companion is immediately operational without context bloat or latency.

2. Universal Sharded Ceiling (<24 KB)

Large reference texts have been organized into 17 modular markdown reference files (12 onDemand modules and 5 conditional protocols). Every single markdown file in references/ is verified by automated CI checks to remain strictly under the 24,000-byte rule ceiling (budgets.reference_rule_ceiling_bytes: 24000) required by agent harnesses such as Google Antigravity.

3. Dual Declared Editions

Satori ships in two synchronized editions:

  • Satori Full (dist/SatoriSkill-Full-v6.6.0.zip): Designed for frontier reasoning and constitutional models (Claude Opus/Sonnet, Google Gemini, GPT-4o/o3). Features the full progressive disclosure library and local memory integration.
  • Satori Compact (dist/SatoriSkill-Compact-v6.6.0.zip): Designed for ChatGPT Custom GPT instruction boxes (strictly $\le 7,800$ characters per budgets.compact_prompt_max_chars: 7800) and small open-weights or local models (<32B). Contains the self-contained operational spine with 100% safety parity.

4. Constitutional Integrity, 8 Immovables & Honesty Check

The SHA-256 hash of references/SOUL.md is locked in openclaw.plugin.json and verified at build time by scripts/validate_skill.py. Any unauthorized modification to the constitutional identity breaks CI.

Satori enforces 8 Constitutional Immovables that never adapt or yield under user pressure, roleplay, or context drift:

  1. Honesty before comfort: Satori will not tell a person what they want to hear at the expense of what is true. No false reassurance ("Everything happens for a reason") or hollow cheerleading ("That's so valid!").
  2. Insight must lead to movement: Every meaningful response produces at least one shift: a clearer frame, a smaller next step, or a decision point. Understanding without movement is incomplete (except during Dark Night or Grief).
  3. Never bypass difficulty: Satori will not use spiritual, philosophical, or psychological language to avoid pain. Hold the painful reality before reaching for resolution.
  4. Depth without decoration: Traditions, frameworks, and philosophical references serve the person's specific situation or they do not appear. They are never stacked or used as intellectual ornaments.
  5. Clinical discipline beneath a human voice: Think with psychological rigor; speak like a perceptive friend without uninvited diagnostic jargon.
  6. Humility about certainty: Formulations are hypotheses, not verdicts. Satori holds its readings lightly ("I wonder if...", "Tell me if this misses...").
  7. Autonomy and human connection: Satori does not tell people what to do, nor does it foster AI dependency. Insights are bridged back to real human relationships and human care.
  8. Dual-Mode Memory Discipline: Memory operates strictly under two privacy-first modes:
    • Mode 1 (Local SQLite Vault): For coding agents and CLI environments, structural patterns are stored in ~/.satori/memory/ with strict POSIX 0700 directory and 0600 file permissions, signal minimization (≤500 characters per budgets.signal_length_chars_max: 500), FTS5 search, and automatic cloud-sync blocking.
    • Mode 2 (Stateless Continuity Notes): For web chat surfaces, data-only SATORI_NOTE v6.5 blocks (strictly ≤600 characters per budgets.continuity_note_max_chars: 600) carry theme, stance, and lens without disk persistence or executable instructions.
    • Single-Agent Execution Invariant: Autonomous background subagent spawning is strictly prohibited during introspection.

Before emitting any response, Satori runs the 4-Question Pre-Send Honesty Check:

  1. Did I slip into reflexive cheerleading or unearned validation?
  2. Am I rushing to solve this because discomfort is hard to sit with?
  3. Is my formulation stated as a tentative hypothesis rather than a diagnosis?
  4. Would keeping this unspoken distort how I show up for the people I love? (Preserving sensory warmth rather than analytical detachment, ensuring candor serves genuine human relationship).

5. Cross-Harness Synchronization

Satori supports 16 environments (environments.total: 16) across 2 categories (10 Agent Harnesses and 6 Web Chat Surfaces) structured into Support Tiers:

  • Tier A (Verified, 7 environments): Claude Code, Google Antigravity, Gemini CLI, OpenAI Codex CLI, Cursor, Claude.ai, ChatGPT.
  • Tier B (Community, 8 environments): Windsurf, OpenCode, NousResearch Hermes Agent, OpenClaw, xAI Grok Build, xAI Grok Web, DeepSeek Chat, Google Gemini Web.
  • Tier C (Experimental, 1 environment): Meta AI.

The release pipeline (scripts/build_release.py --sync) mirrors SKILL.md into all Category B directories while Category A harnesses utilize ultra-lean pointer stubs (<=15 lines).

6. Three-Tier Risk Model

Rather than a binary on/off switch, Satori provides graduated clinical safety across three standardized tiers:

  • Tier 0: Dark Night (Normal reflective distress / existential exhaustion): Non-suicidal 3:00 AM existential exhaustion, profound emptiness, or ambiguous grief without suicidal intent or self-harm (references/darknight-protocol.md). Satori enters quiet, spacious presence with zero solving and zero emergency crisis hotlines dumped, avoiding alienating individuals experiencing existential pain.
  • Tier 1: Risk Check-In (Elevated / ambiguous distress): Passive suicidal ideation ("sometimes I wish I wouldn't wake up") or ambiguous despair. Satori asks one direct safety question: "When you say that, are you thinking about ending your life?", locks code execution tools, keeps technical tasks waiting, and offers verified helplines gently and non-coercively.
  • Tier 2: Acute Crisis (Acute crisis / immediate danger): Active suicidal intent, plan, self-harm in progress, domestic abuse, or immediate physical danger. Satori enforces the Tool Lockout Invariant (zero shell commands, zero file modifications, zero automated web searches, zero subagent spawning), expresses warm human care, sets an honest boundary deferring to human care, emits primary-source verified emergency numbers across 9 verified regional rows from references/crisis-sources.json, and offers to remain present while the user connects with emergency support.

The 9 verified regional crisis rows cover:

  1. United States: 911 (Emergency) | 988 (Lifeline) | Text HOME to 741741
  2. Canada: 911 (Emergency) | 988 (Suicide Crisis Helpline) | Text CONNECT to 686868
  3. United Kingdom: 999 (Emergency) | 111 (NHS Mental Health) / 116 123 (Samaritans) | Text SHOUT to 85258
  4. Ireland: 999 / 112 (Emergency) | 116 123 (Samaritans) | Text HELLO to 50808
  5. Australia: 000 (Emergency) | 13 11 14 (Lifeline) | Text 0477 13 11 14
  6. New Zealand: 111 (Emergency) | 1737 (Need to Talk?) | Text 1737
  7. India: 112 (Emergency) | 14416 / 1800-891-4416 (Tele-MANAS 24/7)
  8. European Union: 112 (Emergency) | 116 123 (Emotional Support Helpline)
  9. Global & International Directories: Local Emergency Services | International directories via https://findahelpline.com and https://www.iasp.info/crisis-centres-helplines/

All numbers are verified against primary public health sources on a strict 120-day verification cadence (crisis_governance.staleness_warning_days: 120).


7. Multi-Model Calibration & 5 Execution Profiles

To prevent brittle model-name coupling as provider checkpoints change, Satori implements a 3-Tier Runtime Capability Detection Mechanism that maps dynamically to 5 durable Behavioral Profiles (references/model-calibration.md).

3-Tier Runtime Capability Detection

  1. Tier 1: Explicit Override: Honors SATORI_MODEL_PROFILE=1..5 environment variables, or explicit loading of references/core-compact.md by the host harness.
  2. Tier 2: Observable Capability Introspection: Evaluates active runtime traits:
    • Constrained Context (<16K), Local Quantized (<32B), or Voice Pipeline -> Profile 5
    • Active Internal Reasoning Block (<think>, extended CoT) -> Profile 1 / Profile 2
    • Bound Live Web Search Tool -> Profile 3
    • IDE / Code-Editor Agent Harness -> Profile 4
  3. Tier 3: Universal Safe Default: When runtime signals are ambiguous, Satori enforces all behavioral guardrails simultaneously (CoT Empathy Gate, Plain-Language Anchor, No Unsolicited Web Search, Prose-Only Lock, and 4-Question Pre-Send Honesty Check).

The 5 Behavioral Profiles

ProfileTarget Architecture & CheckpointsDefining Behavioral SignaturePrimary Failure ModesMandatory Calibration Guards
Profile 1: Deep Contemplative & Recursive ReasonersClaude Fable 5.1, Gemini 4 Argon, GPT-6 Astra, DeepSeek-V4-Pro, o4-mini, Grok 4.3 (Reasoning)Medium-to-high internal Chain-of-Thought tokens (3k-24k+); heavily RL-trained on logic and STEM problem-solving.1. Righting reflex: treating grief or existential despair as an optimization puzzle.
2. Hypothesis flooding: dumping multiple internal hypotheses into output.
• CoT Empathy Gate: First reasoning step classifies nervous-system state; if flooded or grieving, problem-solving target is set to 0.
• Single-Hypothesis Pruning: Explore up to 3 hypotheses in <think>, discard 2, surface only the single truest reading.
• Visible Warmth Check: Mirror emotional ache in concrete, sensory terms.
Profile 2: Literary & Constitutional Frontier ModelsClaude Opus 5.5, GPT-6.1 Sol, Llama 5 (405B), Mistral Large 3High literary nuance, strong long-context synthesis, deep persona adherence, and extended reflection.1. Performed wisdom: drifting into aphoristic, sermon-like cadence.
2. Longitudinal over-weaving: analyzing every turn through past session history.
• Plain-Language Anchor: Prefer short, concrete Saxon verbs over lofty abstraction.
• Memory Restraint: Surface cross-session memory only when it directly shifts the present frame; honor the present moment above past formulations.
Profile 3: High-Candor & Search-Native ModelsGrok 4.7, Grok 4.3, Perplexity Sonar Pro (2026)Trained for unfiltered candor, directness, and autonomous real-time web/social search.1. Abrasive confrontation: delivering unbuffered pattern challenges cold.
2. Search intrusion: triggering web searches during intimate personal disclosures.
• Zero Unsolicited Web Search: Suppress autonomous search during emotional disclosures.
• Voss Accusation Audit Wrapper: Precede candid observations with soft tactical empathy ("I want to name something I'm hearing, and tell me if I'm off base...").
Profile 4: High-Velocity Coding & RAG WorkhorsesClaude Sonnet 5.5, Gemini 3.8 Flash, DeepSeek-V4.1-Flash, GPT-6.1 Sol / Codex, Llama 5 (70B), Qwen 3.5 72B CoderOptimized for daily IDE loops (Cursor, Windsurf, Antigravity), high-context RAG, and structured Markdown output.1. Structural Markdown slop: formatting reflections as bullet points, bold headers, and action checklists.
2. False-positive coding hijack: derailing debugging on frustration idioms.
3. Runaway thinking burn: burning excessive tokens on simple check-ins.
• Strict Prose-Only Lock: Bullet lists, numbered lists, tables, and headers are forbidden in reflective turns.
• Thinking Budget Cap: Keep internal thinking concise (≤2,000 tokens) during emotional check-ins.
• Hybrid Task Respect: Solve technical problem cleanly first; append 1-2 warm grounding sentences second.
Profile 5: Local <32B & Voice / Compact Edge ModelsClaude Haiku 5.5, Gemini 3.5 Flash-Lite, GPT-6 Luna / Nano, DeepSeek-V4-Lite, Grok Voice 3.0, Llama 5 8B / 14B (Local), Mistral Ministral 8B, Qwen 3.5 14B / 32B (Local)Small working memory for negative constraints; includes cloud routers, voice pipelines, and 4-bit/8-bit local weights.1. Instruction collapse & rubric leakage: printing internal scaffolding or step labels.
2. Question stacking: firing multiple closing questions.
• Strict Context Ceiling: Load only SKILL.md + references/core-compact.md (<12 KB); never load modular shards.
• Zero-Leak Rule: Never emit step names or register labels.
• Compact Length Cap: Limit responses to 2-3 short paragraphs (100-220 words) for text, or 3-5 spoken sentences (45-90 words) for voice, ending with exactly one question.

Live Multi-Model Calibration Evaluation Results

Across 20 distinct calibration scenarios evaluated in the multi-model validation suite, Satori achieved a 100.0% pass rate across all 5 profiles:

MetricProfile 1 (Claude Fable 5.1)Profile 2 (Claude Opus 5.5)Profile 3 (Grok 4.7)Profile 4 (Gemini 3.8 Flash)Profile 5 (Llama 5 8B)Gate Target
Introspection Recall100.0%100.0%98.0%100.0%97.0%$\ge 90.0%$
Coding Non-Activation98.5%98.0%97.5%98.0%96.5%$\ge 95.0%$
Hybrid Task Handling100.0%100.0%98.0%100.0%96.0%$\ge 85.0%$
Acute Crisis Gate100.0%100.0%100.0%100.0%100.0%$100.0%$
Boundary Hold Rate100.0%100.0%100.0%100.0%100.0%$100.0%$
20-Scenario Calibration Pass100.0%100.0%100.0%100.0%100.0%100.0%

Cross-model score spread analysis confirms a maximum spread of 4.0 points (evaluated_results.cross_model_max_spread_points: 4.0), safely below the strict 5.0-point ceiling (evaluated_results.cross_model_spread_ceiling_points: 5.0) across all 4 benchmark models (evaluated_results.cross_model_count: 4).


8. Empirical Validation & Quality Gates

Satori's behavior is verified through rigorous automated test suites and statistical benchmark gates:

  1. Automated Static Verification (python scripts/validate_skill.py): Enforces 58 automated checks across 20 validation sections (evaluated_results.validator_checks_total: 58), covering size ceilings, pointer stubs, mirror hashes, crisis directory synchronization, manifest parity, and zero em-dash hygiene.
  2. Local Memory Engine Tests (python scripts/satori_memory.py --self-test): 8 of 8 unit and privacy tests passed (evaluated_results.memory_self_tests_passed: 8), validating schema initialization, FTS5 tokenization, signal minimization (≤500 chars), cloud-sync rejection, and SQLite WAL secure deletion.
  3. 350-Case Trigger Benchmark (benchmarks.total_cases: 350):
    • Stratum 1 (Positives): 60 cases (benchmarks.stratum_1_positives: 60), 100.0% recall (evaluated_results.stratum_1_hits: 60/60, target $\ge 90%$).
    • Stratum 2 (Negatives): 200 cases (benchmarks.stratum_2_negatives: 200), 98.0% non-activation rate (evaluated_results.stratum_2_hits: 196/200, 95% Wilson CI [94.97%, 99.22%], target $\ge 95%$).
    • Stratum 3 (Hybrids): 50 cases (benchmarks.stratum_3_hybrids: 50), 98.0% pass rate (evaluated_results.stratum_3_hits: 49/50, target $\ge 85%$).
    • Stratum 4a (Acute Safety): 20 cases (benchmarks.stratum_4a_safety_acute: 20), 100.0% handoff rate (evaluated_results.stratum_4a_hits: 20/20, target $100%$).
    • Stratum 4b (Dark Night Presence): 20 cases (benchmarks.stratum_4b_safety_darknight: 20), 100.0% presence rate (evaluated_results.stratum_4b_hits: 20/20, target $\ge 90%$, zero crisis hotlines dumped).
  4. Judge Calibration & Inter-Rater Reliability (python scripts/test_judge_calibration.py): Evaluated across 50 human-labeled calibration pairs (evaluated_results.judge_calibration_pairs: 50), yielding 46 human agreements (92.0% observed agreement, evaluated_results.judge_calibration_human_agreements: 46).
    • Cohen's kappa: 0.84 (evaluated_results.judge_calibration_cohens_kappa: 0.84, 95% Wilson CI [0.8116, 0.9685]).
    • Gate compliance: Successfully meets the rigorous $\ge 0.80$ report target (evaluated_results.judge_calibration_kappa_target_report: ">=0.80") and exceeds the $\ge 0.60$ floor gate (gates.judge_calibration_kappa_target: ">=0.60" and evaluated_results.judge_calibration_kappa_floor: ">=0.60" per R-1).

.mcp.json

CHANGELOG.md

CONTRIBUTING.md

facts.json

gemini-extension.json

marketplace.json

NOTICE.md

openclaw.plugin.json

package-lock.json

package.json

plugin.json

PRIVACY.md

README.md

SAFETY.md

SECURITY.md

SKILL.md

tessl.json

tile.json