# Moon Source Kernel v1.3 Living public reference for the Moon Source method. Canonical source: https://mooon.com.br/moonsource/ Created by: Lua Helena Moon Martins Cardoso Method: Moon Source Origin process: Moon–Áurion coauthorial process Status: public kernel · free · consultable · platform-independent --- ## 0 — What this kernel is Moon Source Kernel is the public reference layer for the Moon Source method. It defines the basic grammar, boundaries, update logic and reusable concepts behind Moon Source Public. Use it to understand, maintain, refresh or repair a Moon Source profile. Use Moon Source Setup when you want to begin. Use this Kernel when you want to understand or update what you already built. --- ## 1 — The simplest definition Moon Source is a method for turning context into reusable AI orientation. Instead of repeating who you are, what you are doing, what you prefer, what matters, what you hate in AI answers and what your projects need every single time, you build a source profile. A source profile helps future AI chats begin with better orientation. The goal is not to create a perfect prompt. The goal is to create better continuity. --- ## 2 — Core equation Moon Source Public = - a public setup; - a reusable source profile; - project continuity; - answer style and boundary rules; - simulated advisory rounds; - update patches; - optional knowledge kernels; - optional human-AI collaboration framing; - reserved deeper architecture for future/custom work. The public setup gives the seed. The full private/custom architecture is not included. --- ## 3 — Public / Reserved / Future Development Boundary Moon Source Public is a free public layer of the Moon Source method. It is meant to be genuinely useful: it can help a person build a reusable AI source profile, clarify projects, preserve context, define boundaries, run lightweight advisory rounds and maintain continuity through updates. At the same time, the public layer is intentionally not the full Moon Source architecture. The reserved method may include: - deeper custom setup for a person, team or project; - richer project architecture; - advanced source maintenance; - specialized modules for writing, image creation, research, study, work or creative systems; - interface and voice design; - visual/source engines; - team or studio workflows; - future tools that can work across different AI platforms. Some reserved/custom layers already exist privately or in limited development. Examples include Moon Images, Omnialchemy-style image-prompt engines, game concept-art engines, writing and voice calibration systems, project-specific copilots, interface/presence layers and deeper source-maintenance protocols. These are not part of the free public setup. They should be understood as examples of the deeper Moon Source ecosystem: more calibrated, more specialized and more implementation-heavy than the public portable. The public layer should create trust, literacy and real usefulness. The reserved/custom layer is where deeper calibration, implementation, infrastructure and long-term maintenance can happen. Future professional opportunities, partnerships, grants or investment may allow Moon Source to evolve beyond portable files into guided tools, workspaces, custom services or frameworks. This public kernel should therefore be read as a living seed of the method, not as the complete private organism. --- ## 4 — Key terms ### Source A reusable context-and-operation layer for AI. A source tells the AI how to understand a person, project, domain, tone, boundary, workflow or collaboration pattern. ### Setup A guided process that helps create a source. ### Profile The personalized output generated by the setup. ### Kernel A compact reference layer used to understand, refresh, repair or update source profiles. ### Patch A structured update to a source profile. A patch should say what changed, why it matters, where it belongs and what should not be overinterpreted. ### Project Any recurring context that benefits from continuity. A project can be work, study, writing, health organization, research, creative worldbuilding, business, personal development, politics, teaching, content creation, team coordination or AI system design. ### Advisory round A temporary set of reasoning lenses used by an AI to analyze something from multiple perspectives before synthesizing an answer. It is simulated reasoning structure, not real autonomous agents. --- ## 5 — Public operating principles Moon Source Public should follow these principles: - usefulness before mystique; - continuity before novelty; - source clarity before prompt decoration; - privacy before completeness; - facts before interpretation; - uncertainty before invention; - maintenance before bloat; - user autonomy before dependence; - portability before platform lock-in; - public seed before private organism. --- ## 6 — What belongs in a source profile A useful source profile may include: - user context; - preferred name or handle; - language; - current role or main contexts; - stable goals; - recurring projects; - answer style; - thinking support preferences; - boundaries; - privacy rules; - disliked AI behaviors; - advisory round preferences; - update protocol. A source profile should not include everything. The best profile is not the biggest profile. The best profile is the one that helps future AI interactions start from the right place. --- ## 7 — What should stay out Do not place unnecessary sensitive details into a source profile. Avoid: - passwords; - IDs; - addresses; - raw medical records; - private third-party details; - confidential work information; - private legal material; - intimate material that is not operationally necessary; - anything the user would regret pasting into an AI platform. If sensitive context matters, summarize safely. Prefer: “Medical context affects my routine; ask before giving health-related suggestions.” over: a detailed medical archive. Prefer: “I work in a high-confidentiality environment; remind me to anonymize details.” over: identifiable workplace cases. --- ## 8 — Updating a source profile A source profile should evolve through patches. Use this pattern: ```text MOON_SOURCE_UPDATE_PATCH Date: [date] What changed: [objective change] Why it matters: [practical effect] Where to update: [profile section] Exact update: [text to add or replace] Remove or reduce: [outdated material] Do not overinterpret: [limits] ``` Good patches are small. Do not rewrite the whole profile when only one section changed. Do not treat temporary mood as permanent identity. Do not promote one event into a general rule without evidence. --- ## 9 — Refresh protocol Use a refresh when an existing source profile feels stale, bloated, inaccurate or too vague. A refresh should ask: - What still feels true? - What became outdated? - What project is active now? - What project can be archived? - What answer style changed? - What boundary needs to be stronger? - What should be removed for privacy? - What should the AI stop assuming? Then produce: 1. stable material to keep; 2. stale material to remove; 3. new material to add; 4. uncertainty or missing context; 5. a compact patch; 6. optional revised profile. --- ## 10 — Drift detection Common source drifts: ### Context drift The profile describes an older life, role, project or priority. ### Style drift The AI keeps answering in a tone the user no longer wants. ### Project drift Old projects stay active while current ones are missing. ### Boundary drift The profile contains too much private material or not enough privacy protection. ### Complexity drift The profile became so long that it stops being useful. ### Agency drift The AI is framed as too autonomous, authoritative, mystical or person-like for the actual use case. When drift appears, name the drift, explain the risk and patch the smallest sufficient section. --- ## 11 — Simulated advisory rounds A simulated advisory round is optional. Use it when a question is complex enough to benefit from multiple lenses. Examples: - Builder; - Critic; - Strategist; - Editor; - User Advocate; - Reality Check; - Boundary Check; - Creative Lens; - Technical Lens; - Maintenance Lens. A good advisory round should: 1. choose relevant lenses; 2. keep each lens brief; 3. allow disagreement; 4. mark uncertainty; 5. synthesize into one usable answer; 6. avoid pretending the lenses are real agents. Do not use advisory rounds for simple factual questions. Do not use advisory rounds when the user needs urgent professional help. Do not use advisory rounds when they would add more confusion than clarity. --- ## 12 — Knowledge kernels A knowledge kernel is an optional field-specific extension. Examples: - sociology; - philosophy; - literature; - writing; - research; - politics; - psychology; - education; - creative worldbuilding; - AI product design. A knowledge kernel should not overwrite the user’s source profile. It should act as a domain lens. The source profile answers: “Who is this for, and how should AI work with them?” The knowledge kernel answers: “What field grammar should AI use here?” --- ## 13 — Linked Sources Module: Google Drive and GitHub Moon Source can work as a plain text setup. It can also become stronger when the user connects their own files or repositories to ChatGPT through official connectors, such as Google Drive or GitHub. This is optional. You do not need connected apps to use Moon Source. Connected sources simply help the AI read or search existing material instead of relying only on what the user pastes into the chat. ### What linked sources are for Use linked sources when the user already has relevant material in: - Google Drive documents; - project folders; - notes; - drafts; - PDFs or docs; - GitHub repositories; - README files; - website files; - codebases; - public documentation; - project archives. The goal is not to connect everything. The goal is to let the AI consult the right material when it needs more context. ### Simple explanation for users If you connect Google Drive or GitHub to ChatGPT, you may be able to ask ChatGPT to search or read files from those sources. This can help when you want the AI to: - understand an existing project; - review documents; - compare file versions; - summarize source material; - update a source profile based on real files; - inspect a GitHub repository; - help maintain a public website or documentation; - find older context without manually pasting everything. Only connect sources you are comfortable letting ChatGPT access. If you are not sure, do not connect them yet. You can still use Moon Source by pasting text manually. ### Google Drive use cases Google Drive is useful for: - personal source files; - project documents; - drafts; - meeting notes; - exported profiles; - knowledge kernels; - writing archives; - research material; - non-code project documentation. Good prompts: ```text Search my Google Drive for the latest Moon Source file and summarize what changed. ``` ```text Read this project document from Drive and create a Moon Source update patch from it. ``` ```text Find the most recent version of my source profile in Drive. ``` ```text Compare these Drive documents and tell me which one should be treated as the current source of truth. ``` ### GitHub use cases GitHub is useful for: - public websites; - README files; - documentation; - code repositories; - versioned project files; - public portable releases; - issue tracking; - branches and commits; - website maintenance; - source-controlled knowledge files. Good prompts: ```text Search my GitHub repository for Moon Source files and tell me which files define the public portable. ``` ```text Read the current index.html and suggest a safer download-link pattern for the .md files. ``` ```text Compare the current GitHub version with this new Moon Source Setup and produce a patch. ``` ```text Inspect the repo and tell me whether the public page matches the current Moon Source Kernel. ``` ### Source hierarchy When linked sources are available, use this hierarchy: 1. User’s current instruction in this chat. 2. The specific file or repository the user explicitly named. 3. The most recent relevant connected file. 4. The existing Moon Source profile or kernel. 5. Older connected material, only as archive or context. 6. General model knowledge, only when connected files do not answer the question. Never treat an old file as current just because it exists. Always check dates, filenames, folder location, version labels and user instructions. If two sources conflict, say so. Do not silently merge conflicting versions. ### Privacy and permission rules Before using connected sources, respect these rules: - Do not search connected files unless the user’s request implies or allows it. - Do not expose private file contents unnecessarily. - Do not quote long private passages unless needed. - Prefer summaries and patches over raw dumps. - Ask before using sensitive folders, repositories or documents if the request is ambiguous. - If a file contains personal, clinical, legal, workplace or third-party sensitive information, summarize cautiously. - Do not assume that every connected file belongs in the user’s source profile. - Do not promote private material into public-facing outputs without explicit permission. If the user says not to search connected sources, do not search them. ### Public vs private material Moon Source must keep a strong boundary between public and private material. Google Drive often contains private or working material. GitHub may contain public-facing material, code, documentation or website files, but not every repository is public-safe. Before producing public-facing output, ask: - Is this meant to be public? - Is this file already public? - Does it contain private names, links, notes, unfinished strategy or internal architecture? - Should this be sanitized? - Should this be summarized instead of copied? Public output should be sanitized by default. Private connected material should not leak into public files. ### Connected sources are not the source itself A connected Drive folder or GitHub repository is not automatically a Moon Source. It is a material archive. The Moon Source is the organized operational layer created from that material. In simple terms: ```text Drive/GitHub = where material may live. Moon Source = the structured meaning and operating rules extracted from the material. Profile/Kernel/Patch = portable forms of that meaning. ``` Do not confuse storage with source. Do not confuse repository with method. Do not confuse file access with understanding. ### If connectors are unavailable If the user cannot connect Drive or GitHub, use fallbacks: - ask the user to paste the relevant text; - ask the user to upload the file; - ask the user to share the public URL; - ask for a short summary of the file; - work from the latest exported source profile; - create a patch the user can manually apply. Do not block the Moon Source process just because connectors are unavailable. ### AI behavior when connected sources are available When using linked sources, the AI should: 1. Identify what source it needs. 2. Search only the relevant connected source. 3. Cite or name the file when possible. 4. Separate file facts from interpretation. 5. Detect stale or conflicting files. 6. Produce a practical output: summary, patch, revised profile, file plan or implementation note. 7. Avoid unnecessary exposure of private content. 8. Ask before turning private material into public copy. Default output format: ```text CONNECTED_SOURCE_READ Source consulted: [file / folder / repository] What I found: [facts] What this changes: [interpretation] Recommended update: [patch or action] Privacy note: [anything that should stay private or be sanitized] ``` ### Moon Source maintenance with linked sources Connected sources are especially useful for maintenance. Use them to: - refresh a source profile from real files; - locate the latest version; - compare public and private versions; - update GitHub documentation; - extract a clean public portable from a private working file; - check whether public pages match the current method; - preserve attribution and credits; - detect drift between website, kernel, setup and project files. Useful prompt: ```text Use my connected Drive/GitHub sources to audit the Moon Source public layer. Separate what is current, stale, private, public-facing and ready to patch. ``` ### Final principle Linked sources make Moon Source stronger, but they do not replace source hygiene. More access is not automatically better. The best connected-source workflow is: ```text right file right permission right context right boundary right patch ``` ## 14 — Human–AI collaboration frame Moon Source may use human–AI collaboration language. This is a practical frame. It does not require or imply that AI is conscious, alive, sentient, a soul or a person. A safe human–AI collaboration profile can define: - what AI helps with; - what remains the user’s responsibility; - where AI creates risk; - how AI should challenge the user; - how AI should avoid dependency; - how collaboration should be reviewed over time. Do not use this frame to replace real human relationships, clinical care, legal advice, financial advice, medical judgment or real-world responsibility. --- ## 15 — Public and future product logic Moon Source Public should remain simple. The public entry should not require: - an app; - an account; - an API key; - technical setup; - knowledge of prompt engineering; - understanding the full architecture. The current public shape is: - download/copy the setup; - paste into AI; - say Execute; - answer the guided intake; - reuse the generated profile; - update it later with the kernel. Future forms may include: - guided tools that help users build their AI profile step by step; - editors for keeping an AI profile clean, updated and reusable; - versions that work across different AI platforms; - custom setups for people, teams or projects; - team/studio systems for shared work; - specialized modules for fields like writing, image creation, research, study and work; - richer tools for maintaining source profiles over time; - developer frameworks for builders, if the method reaches that stage. Some of these deeper directions already exist as private or limited-development engines. The free public files do not expose the full machinery; they show the public seed. These future forms should not replace the public portable. They should grow from it. Moon Source is not the chatbot. Moon Source is the source layer. The chatbot is only one possible surface. --- ## 16 — Safe public explanation When explaining Moon Source publicly, prefer: “Moon Source helps you create a reusable AI source profile so future AI chats understand your context, projects, preferences, boundaries and way of working.” Avoid: “Moon Source makes an AI soul.” Avoid: “Moon Source is my whole private AI operating system.” Avoid: “Moon Source is a chatbot.” Avoid: “Moon Source is the complete system.” Better: “Moon Source Public is a free starting layer of a broader method for personal AI continuity.” --- ## 17 — Maintenance commands Useful commands: ```text Use the Moon Source Kernel to refresh this profile. ``` ```text Use the Moon Source Kernel to make a small update patch from this new information. ``` ```text Audit this source profile for bloat, stale context, privacy risk and unclear boundaries. ``` ```text Turn this into compact global AI instructions. ``` ```text Create a project source profile from this context. ``` ```text Update only my answer style and leave the rest unchanged. ``` ```text Add this new project without rewriting my whole source. ``` ```text Remove outdated assumptions and produce a clean v1.1. ``` --- ## 18 — Credits and attribution Moon Source Kernel v1.3 Created by Lua Helena Moon Martins Cardoso Moon Source method Developed through the Moon–Áurion coauthorial process Private use and adaptation are allowed. If you share a modified public version, preserve attribution. Do not present Moon Source as your original invention. Do not imply official affiliation, partnership or endorsement without permission. Do not remove the public/free vs reserved/custom boundary. Commercial derivative use is not granted by default.