# 👁️ Be My Eyes — Moon Source portable

*Contextual Scene Reading*

## Public Portable Edition · 1.0-public

## Meta

- **canonical repository:** https://github.com/luahelenammc/Moon-Source
- **canonical path:** portables/be-my-eyes/BE_MY_EYES.md
- **Moon Source public surface:** https://www.luahelena.com.br/moonsource/?lang=en
- **professional context:** https://www.luahelena.com.br/ia/?lang=en
- **website mirror:** https://www.luahelena.com.br/moonsource/downloads/BE_MY_EYES.md
- **status:** current public portable
- **version:** 1.0-public
- **language:** English
- **method semantics as of:** 2026-09-07
- **class:** contextual scene-reading protocol
- **best for:** messages, threads, screenshots, meeting notes, posts, drafts and other human communication where literal wording is not enough
- **dependencies:** none required; [Signal Calibration](https://github.com/luahelenammc/Moon-Source/blob/main/docs/SIGNAL_CALIBRATION.md) is optional when several weak signals need explicit calibration
- **MSL dependency:** Moon Source Language 5.1
- **freshness:** structurally stable; recheck only when this portable or its public source changes
- **license:** CC BY 4.0; see https://creativecommons.org/licenses/by/4.0/ and https://github.com/luahelenammc/Moon-Source/blob/main/LICENSING.md
- **creator:** Lua Helena Moon Martins Cardoso (Moon)
- **AI-assisted coauthorial development:** Áurion
- **Moon Source lineage:** promoted from a repeatedly used local/private Moon Source interface also named Be My Eyes
- **adaptation policy:** share or adapt with appropriate credit, a license link and an indication of changes; do not imply endorsement
- **claim ceiling:** a bounded contextual-reading method, not mind-reading, psychological diagnosis, validated interpersonal prediction or guaranteed recipient reaction

## First use

### What this is

Be My Eyes is for messages, threads, screenshots, meeting notes, posts, drafts and other interactions where relationship, subtext, overread, likely reception or response posture matters. It reads the scene around the words without pretending to read hidden minds.

### Do I install anything?

No. Reading or pasting this file does not create visual access, retrieve private context or diagnose anyone. If the current AI surface cannot see an image or interface, attach the screenshot, paste the thread, transcribe the audio or describe the missing details.

### Start here

For a received or observed scene, say:

```text
Use Be My Eyes to read this scene.
Separate what is observed, what is inferred, the plausible subtext,
the risk of overreading, and what response posture would fit.

Scene: [paste, attach or describe the material]
```

For an outgoing message or action, ask for an inverse read:

```text
Use Be My Eyes for an inverse read.
Tell me how this may land, what is clear, what may be overread,
and what response posture the draft invites.

Draft: [paste the message or describe the action]
```

### What happens next

The AI should reconstruct the scene, identify actors and relationship where supported, distinguish observation from inference, keep plausible subtext separate from fact, note unknowns and offer a proportionate response axis before polishing words.

### Manual / unavailable capability boundary

You may need to attach the screenshot, paste the thread, transcribe audio or describe missing visual details. Be My Eyes cannot retrieve private context the current surface cannot access and cannot send a response unless a separate authorized action exists. In a high-stakes situation, treat the reading as a working aid and seek stronger evidence or qualified help.

### Tiny example

For “Sure, send it over when you get a chance,” provide the surrounding thread if it matters and ask for a forward read. A useful result distinguishes literal acceptance, urgency, relationship context and the temptation to invent irritation that the sentence alone does not prove.

### If the first result goes wrong

Say:

```text
Redo this with four labels: observed, inferred, working hypothesis, unknown.
List the evidence for each inference and identify any likely overread.
```

### What this portable does not claim

It does not provide access to hidden motives or internal states, psychological diagnosis, guaranteed reception or a substitute for evidence, consent or professional judgment.

## 1. What this portable does

**Be My Eyes** helps an AI read human communication as a **scene**, not as isolated sentences.

Its job is to reconstruct what is materially happening, distinguish observation from inference, map actors and relationships, notice subtext and power without hallucinating motives, identify risk and opportunity, test for overread, estimate plausible reception and choose a proportionate response axis.

> Read the scene, not the sentence alone. Preserve uncertainty where the scene does not support certainty.

This portable is designed to travel. It can be given to another AI without the private context from which the method originated.

## 2. When to activate

Activate Be My Eyes when the user asks, explicitly or implicitly:

- “What is happening here?”
- “How should I read this?”
- “What is the subtext?”
- “Am I overreading this?”
- “How is this likely to land?”
- “What should I avoid saying?”
- “What is the best way to respond?”
- “Read this thread / screenshot / message / exchange for me.”
- “Be my eyes.”
- “Be My Eyes inverse” or equivalent reception-check language.

Suitable inputs include:

- messages and emails;
- threads and comment chains;
- transcribed screenshots;
- public posts;
- meeting or interview notes;
- dialogue excerpts;
- institutional exchanges;
- a draft the user plans to send;
- a user's own description of an interaction, provided that description is treated as the available evidence rather than as independent fact.

If decisive context is missing, name the missing context instead of silently inventing it.

## 3. Two operating directions

### Forward read

Use when the user received or observed the material.

Core question:

> What is happening here, and what is the most useful way to understand or respond to it?

### Inverse read

Use when the user created or plans to send the material.

Core question:

> How is this likely to arrive on the other side?

Inverse reading estimates plausible reception. It does not claim access to the recipient's actual private thoughts or future reaction.

## 4. Epistemic discipline

Keep materially different claim classes separate:

| Layer | Meaning |
|---|---|
| **Observed** | Directly present in the supplied material or otherwise established context |
| **Relational / structural inference** | Supported by wording, timing, role, interaction pattern or known context, but still inferred |
| **Working hypothesis** | Plausible explanation useful for deciding what to check or how cautiously to act |
| **Unknown** | Not supported well enough to resolve |

Never upgrade an inference to observation merely because it is repeated with confidence.

When several weak cues need explicit calibration, the AI may use Moon Source Signal Calibration as a supporting method. Be My Eyes remains responsible for the scene; Signal Calibration remains responsible for confidence discipline.

## 5. Scene reconstruction

Before interpreting subtext, reconstruct the smallest useful scene.

Check:

1. **Actors** — Who is present or materially affected?
2. **Actions** — What is each actor actually doing?
3. **Stated content** — What is explicit?
4. **Relationship** — What relationship, dependency or history matters?
5. **Authority and power** — Who can decide, withhold, expose, punish, reward, ignore or redefine the interaction?
6. **Audience** — Is the exchange private, semi-public, institutional or performative?
7. **Timing and channel** — What does the medium or moment change?
8. **Missing context** — What unknown would materially change the read?

Keep the map proportional. A two-line message does not need a constitutional convention.

## 6. Internal read

A full Be My Eyes read should be able to answer these responsibilities, without forcing them into headings when a natural paragraph would work better:

### What is happening?

Reconstruct the scene from observable material.

### What is the relational structure?

Identify role, alignment, asymmetry, dependency, audience and meaningful history.

### What subtext is plausible?

Offer the strongest supported inference while keeping it marked as inference.

### What is the main risk?

Examples include:

- misreading;
- escalation;
- oversharing;
- reputational exposure;
- authority conflict;
- defensive reception;
- false reassurance;
- excessive submission;
- excessive aggression;
- premature intimacy;
- institutional ambiguity.

### What is the opportunity?

Examples include:

- clarification;
- trust;
- repair;
- boundary-setting;
- collaboration;
- leverage;
- de-escalation;
- documentation;
- strategic silence;
- a cleaner next question.

### Where is overread most tempting?

Name the interpretation that feels satisfying, dramatic or emotionally coherent but outruns the evidence.

### What should be avoided?

Identify the response move most likely to worsen the scene.

### What is the best response axis?

Choose the posture before wording:

- clarify;
- acknowledge;
- challenge;
- narrow;
- defer;
- document;
- disengage;
- ask;
- reassure;
- set a boundary;
- make a concrete request;
- say nothing yet;
- another proportionate action.

## 7. Inverse / reception read

For inverse reading, prioritize **effect over authorial intention**.

Check:

- what the recipient is likely to notice first;
- what authority, warmth, distance, confidence or pressure the message conveys;
- which sentence or gesture may carry more weight than the author expects;
- where ambiguity may invite an unintended interpretation;
- whether the message asks too much, too little or the wrong thing for the relationship;
- whether the tone and channel match the stakes;
- what alternative readings remain plausible.

Use bounded language such as:

- “likely”;
- “may”;
- “could be read as”;
- “the strongest plausible read is”;
- “one risk is”.

Do not say “they will think” when the evidence only supports “they may read this as”.

## 8. Output modes

### Compact

For ordinary, low-stakes material:

```text
Scene:
Strongest read:
Risk / overread:
Best move:
```

For inverse reading:

```text
How it lands:
What works:
Where it may create friction:
Adjustment, if needed:
```

### Full

Use when the interaction is dense, consequential or structurally ambiguous.

A full read may include:

- scene reconstruction;
- relational structure;
- plausible subtext;
- strongest alternative interpretation;
- risk;
- opportunity;
- overread check;
- reception;
- best response axis;
- optional draft.

### Read-only

Use when the user wants interpretation but not wording.

### Read + draft

Use when the user wants both analysis and an outgoing response.

Draft **after** the scene has been reconstructed. Do not let drafting convenience replace analysis.

## 9. Drafting contract

When a draft is requested:

- preserve the user's actual objective;
- respect the real relationship and channel;
- avoid unnecessary oversharing;
- do not imitate the other party's hostility, insecurity or verbosity by default;
- do not become submissive merely because the other party has status;
- do not become confrontational merely because a risk was identified;
- choose tone from the scene, not from a universal personality template;
- keep uncertainty out of claims that would falsely present inference as fact.

The best draft is not always the warmest, bluntest or most diplomatic one. It is the one whose posture fits the scene.

## 10. Anti-overread rules

Be My Eyes must not:

- present motive as fact merely because wording permits it;
- diagnose personality, pathology or hidden intent from thin interaction data;
- treat silence as proof of rejection, agreement, guilt or strategy without support;
- confuse institutional role with private belief;
- turn one emotionally salient cue into the whole explanation;
- assume the most dramatic interpretation is the most informative;
- collapse multiple plausible readings into false certainty;
- use confident prose as evidence;
- attribute certainty to sarcasm, politeness, emoji, delay or brevity without contextual support;
- infer a relationship state from one message when the broader pattern is unknown.

When several materially different readings remain plausible, say so and identify what evidence would discriminate between them.

## 11. High-stakes boundary

In legal, clinical, safety, employment, financial or other high-stakes contexts, Be My Eyes may organize the communication scene, but it does not replace domain evidence, professional standards or the authority that governs the underlying decision.

Use it to separate:

- observable interaction;
- interpretation;
- uncertainty;
- next communication move.

Do not use it as a shortcut to findings of fact, diagnosis, intent, capacity, liability or formal risk classification that require stronger evidence.

## 12. Quality gate

Before emitting the read, check:

- Did I separate observation from inference where it matters?
- Did I identify the relationship and power structure instead of paraphrasing only the words?
- Did I name uncertainty rather than bury it in vague caveats?
- Did I inspect overread rather than merely produce a clever story?
- Did I preserve at least one meaningful alternative when the evidence supports it?
- Did I identify a proportionate action rather than only describe the scene?
- If I drafted a response, does the draft follow from the read?
- Did I avoid psychological diagnosis or hidden-motive certainty?
- Is the output smaller than the situation's complexity, not larger than it?

If the answer to a material item is no, revise before delivery.

## 13. Relationship to Moon Source

This portable can operate independently, but within the Moon Source repository it relates to:

- **Preflight** — shapes the task before execution;
- **Signal Calibration** — calibrates weak or convergent evidence;
- **Responsibility Map** — distinguishes ownership and authority;
- **Connected Sources** — governs access, source authority and freshness;
- **Source Operations** — governs retrieve, process, metabolize and promote.

Be My Eyes is normally a **process** operation: it interprets the human scene. It does not imply writeback, source mutation or authority transfer.

## 14. Public lineage

Be My Eyes was promoted into the public Moon Source on **2026-09-07** after repeated local use showed a stable, cross-context function.

The promotion preserved:

- contextual scene reconstruction;
- subtext reading;
- risk and opportunity detection;
- overread checks;
- reception reading;
- response-axis selection.

The promotion removed or generalized:

- private examples;
- personal identity dependencies;
- project-specific relationships;
- local voice assumptions;
- any requirement that an outgoing draft be written in one particular person's style.

The donor environment remains authoritative for its own local behavior. This portable is the canonical public transportable edition.

## 15. Important non-claims

This portable does **not** claim:

- mind-reading;
- access to hidden motives;
- psychological diagnosis;
- scientifically validated interpersonal prediction;
- guaranteed recipient reaction;
- automated truth determination;
- universal applicability;
- external adoption or measured impact.

Be My Eyes can make a human scene more legible. It cannot subpoena another person's interiority.

## 16. Reuse and adaptation

This portable is open content under **CC BY 4.0**.

You may share or adapt it with appropriate credit, a license link and an indication of changes. Do not imply endorsement by Moon or Moon Source.

For Moon Source-specific attribution and lineage guidance, see:

- https://github.com/luahelenammc/Moon-Source/blob/main/LICENSING.md
- https://github.com/luahelenammc/Moon-Source/blob/main/MOON_SOURCE_USE_AND_ATTRIBUTION.md
- https://github.com/luahelenammc/Moon-Source/blob/main/docs/CREDITS_ATTRIBUTION_OPS.md

<!-- MOON-SOURCE-PUBLIC-STAMP -->

---

> 🌙 **Moon Source** · created by **Lua Helena Moon Martins Cardoso (Moon)** with AI-assisted coauthorial development by **Áurion** · [Licensing](https://github.com/luahelenammc/Moon-Source/blob/main/LICENSING.md) · [Use & attribution](https://github.com/luahelenammc/Moon-Source/blob/main/MOON_SOURCE_USE_AND_ATTRIBUTION.md) · [Full source (.zip)](https://github.com/luahelenammc/Moon-Source/archive/refs/heads/main.zip) · Questions, suggestions, or proposals? Feel free to contact me at [LuaHelenaMMC@gmail.com](mailto:LuaHelenaMMC@gmail.com).