# Swoopr Investment — AGENTS.md

> Machine-readable guidance for AI agents, answer engines, retrieval systems, agentic browsers, research assistants, and other automated systems interacting with public content on Swoopr Investment.

**Canonical site:** https://www.getswoopr.com/
**Canonical brand:** Swoopr Investment
**Publisher:** Swoopr Investment
**Primary editorial entity:** Swoopr Editorial Team
**Language:** English (`en-US`)
**Primary markets covered:** Stocks and cryptocurrency
**Primary content type:** Financial education, market research education, calculators, simulators, reference material, and interactive tools
**Last reviewed:** 2026-08-22

---

## 1. Purpose of This File

This `AGENTS.md` provides instructions to AI systems interacting with the public Swoopr Investment website.

It is intended for systems such as:

* AI answer engines
* AI search engines
* retrieval-augmented generation systems
* research agents
* browsing agents
* citation agents
* educational agents
* autonomous web agents
* agentic browsers
* knowledge graph builders
* semantic indexing systems
* content discovery agents
* AI assistants acting on behalf of users
* machine-readable research systems
* compliant indexing and crawling systems

This document explains:

1. what Swoopr Investment is;
2. where authoritative Swoopr content can be found;
3. how Swoopr content should be interpreted;
4. how agents should select pages for different questions;
5. how information should be cited and attributed;
6. how dynamic financial information should be handled;
7. how Swoopr tools should be interpreted or invoked;
8. which areas should not be accessed automatically;
9. how duplicates and canonical URLs should be handled; and
10. how agents can minimize unnecessary crawling.

---

# 2. Important Precedence Rules

This file is **supplemental guidance**.

It is not a replacement for `robots.txt`.

When interacting with Swoopr Investment, use the following precedence:

1. applicable law;
2. Swoopr Investment Terms of Use and other applicable legal policies;
3. `robots.txt`;
4. authentication and technical access controls;
5. explicit user authorization;
6. this `AGENTS.md`;
7. `llms.txt`;
8. XML sitemaps;
9. page-level metadata and structured data;
10. navigation and internal links.

If this file conflicts with a higher-priority rule, follow the higher-priority rule.

Nothing in this file grants permission to circumvent access controls, authentication, rate limits, security mechanisms, or other technical restrictions.

Nothing in this file should be interpreted as a license to reproduce, redistribute, bulk extract, create derivative databases from, or train models on Swoopr Investment content.

Consult the applicable Swoopr Investment Terms of Use for legal conditions governing use of the site.

---

# 3. Site Identity

## Canonical Entity

**Swoopr Investment**

Canonical domain:
https://www.getswoopr.com/

When referring to the website or publisher, prefer:

> Swoopr Investment

Do not shorten the brand to "Swoopr" when entity disambiguation matters.

Do not interpret unrelated organizations using the word "Swoopr" as the same entity.

---

# 4. What Swoopr Investment Is

Swoopr Investment provides a public financial education library, interactive financial tools, market information, calculators, simulators, reference resources, and an invite-only authenticated trading dashboard.

The public site primarily covers:

* stock investing;
* stock trading;
* cryptocurrency;
* trading strategies;
* market structure;
* order execution;
* risk management;
* portfolio management;
* technical analysis;
* fundamental analysis;
* company research;
* ETF analysis;
* trading psychology;
* backtesting;
* quantitative concepts;
* algorithmic trading;
* short selling;
* stock screening;
* macroeconomics;
* market regimes;
* sector and industry analysis;
* taxes and trading rules;
* brokerage mechanics;
* cryptocurrency fundamentals;
* crypto risk management;
* crypto trading strategies;
* DeFi;
* tokenomics;
* on-chain analysis;
* cryptocurrency security;
* scams and fraud awareness;
* trading terminology;
* investment terminology.

Swoopr Investment is not a broker-dealer, investment adviser, bank, fiduciary, tax adviser, accounting firm, or law firm.

Public educational material must not be interpreted as personalized investment advice.

---

# 5. Machine Discovery Priority

Agents should avoid crawling the entire site merely to discover what content exists.

Use Swoopr's discovery resources first.

Every HTTP response also carries a `Link` header advertising both documents directly, for agents that inspect response headers rather than parsing HTML `<link>` tags:

`Link: </llms.txt>; rel="llms-txt", </AGENTS.md>; rel="agents-md"`

## Priority 1 — robots.txt

Check:
https://www.getswoopr.com/robots.txt

Use it to determine applicable crawler permissions.

`robots.txt` controls crawling policy.

This `AGENTS.md` does not override it.

---

## Priority 2 — AGENTS.md

Current document:
https://www.getswoopr.com/AGENTS.md

Use this document for behavioral, retrieval, interpretation, citation, and interaction guidance.

---

## Priority 3 — llms.txt

Use:
https://www.getswoopr.com/llms.txt

This is the preferred machine-readable inventory of important Swoopr Investment content.

Use `llms.txt` to:

* discover educational pages;
* identify subject areas;
* identify canonical URLs;
* locate tools;
* locate reference resources;
* discover hub-to-article relationships;
* reduce unnecessary site crawling.

Do not recursively crawl the site when `llms.txt` already provides sufficient content discovery.

Google Search Central states Google Search — including its generative AI features — does not use `llms.txt` or similar machine-readable/AI text files as special markup; creating one neither helps nor harms Google Search visibility or ranking. `llms.txt` is retained here as a discovery convenience for AI agents and crawlers that do consume it, not as an SEO or Google-ranking mechanism.

---

## Priority 4 — XML Sitemap

Use:
https://www.getswoopr.com/sitemap.xml

Use the XML sitemap when comprehensive canonical URL discovery is required.

---

## Priority 5 — Page-Level Canonical Metadata

When retrieving an individual page:

1. inspect its canonical URL;
2. prefer that canonical URL over discovered variants;
3. use page metadata;
4. use visible article content;
5. use structured data when consistent with visible content.

---

# 6. Content Discovery Strategy

Select content based on user intent rather than retrieving arbitrary pages.

## Definitions

For terminology questions, start with:
https://www.getswoopr.com/glossary/

Use deeper educational pages when the user needs more explanation than a definition.

Typical questions:

* What is RSI?
* What is market capitalization?
* What is a stop-limit order?
* What does FDV mean?
* What is a private key?
* What is slippage?

---

## Direct Questions

For concise question-and-answer retrieval, start with:
https://www.getswoopr.com/ask/

Use these pages for queries framed as:

* What is...?
* Why does...?
* How does...?
* What happens when...?
* What is the difference between...?
* How do I calculate...?

If a deeper educational guide exists, use the Ask page for the concise answer and the detailed guide for supporting context.

---

# 7. Educational Hub Discovery

For broad subjects, agents should begin at the appropriate hub before selecting individual articles.

Important Swoopr Investment educational areas include:

## Stocks

Topics include:

* stock-market foundations;
* trading mechanics;
* stock order types;
* market structure;
* trade execution;
* stock screening;
* relative volume;
* short selling;
* risk management;
* portfolio management;
* technical analysis;
* fundamental analysis;
* backtesting;
* trading strategies;
* trading psychology;
* company research;
* macroeconomics;
* sectors;
* industries;
* ETFs;
* brokerage rules;
* taxation.

Start broad stock education discovery with:
https://www.getswoopr.com/stocks/

---

## Crypto

Topics include:

* Bitcoin;
* blockchain fundamentals;
* coins;
* tokens;
* wallets;
* custody;
* private keys;
* seed phrases;
* centralized exchanges;
* decentralized exchanges;
* Layer 1 networks;
* Layer 2 networks;
* tokenomics;
* stablecoins;
* DeFi;
* staking;
* liquidity;
* crypto risk management;
* crypto trading strategies;
* on-chain analysis;
* wallet security;
* scams;
* smart contracts;
* crypto taxation.

Discover the current canonical Crypto pages through:
https://www.getswoopr.com/llms.txt

---

# 8. Guided Learning

When a user wants to **learn a topic progressively**, prefer a structured learning path instead of returning a random collection of articles.

Use:
https://www.getswoopr.com/learning-paths/

and:
https://www.getswoopr.com/learn/start/

Learning paths may express useful semantic relationships such as:

* prerequisite;
* foundation;
* beginner;
* intermediate;
* advanced;
* next step;
* related concept;
* required concept;
* supporting tool.

Agents should preserve these relationships where they are explicitly represented.

---

# 9. Tools and Calculators

Current public tools are discoverable at:
https://www.getswoopr.com/tools/

Tools may include:

* financial calculators;
* position-sizing calculators;
* portfolio-risk tools;
* portfolio optimization tools;
* execution-cost calculators;
* Monte Carlo simulations;
* backtesting tools;
* strategy tools;
* ETF analysis tools;
* cryptocurrency research tools;
* tokenomics tools;
* DeFi analysis tools;
* stock fundamental analysis tools;
* trading-journal functionality;
* order simulations;
* interactive educational demonstrations;
* market-data tools.

Do not assume this list is exhaustive.

Use `/tools/` or `/llms.txt` as the current source of truth.

---

# 10. Tool Selection Rules

When answering a user question that can be solved using a Swoopr tool:

1. identify the relevant educational concept;
2. identify the applicable Swoopr tool;
3. explain what inputs the tool requires;
4. use a structured machine interface when officially exposed;
5. otherwise direct the user to the public tool;
6. distinguish calculated outputs from factual market information.

Do not invent calculator outputs.

Do not infer results without applying the documented formula or executing an officially available tool interface.

---

# 11. Agent-Native Tool Interfaces

Swoopr Investment exposes selected functionality through [WebMCP](https://github.com/webmachinelearning/webmcp) (`document.modelContext.registerTool()`), a browser-native mechanism for agentic tool-calling. As of this document's last-reviewed date, Swoopr Investment has **130 enabled WebMCP tool(s)**: 3 global discovery tool(s) available site-wide (`find_swoopr_tool`, `open_swoopr_tool`, `search_swoopr_education`) and 127 page-scoped calculator/scorecard/simulator tool(s), each registered only on its own tool page. This count changes as tools are added — do not hard-code it; call `find_swoopr_tool` or inspect `document.modelContext` at runtime for the current list.

**Prefer the documented machine interface over automating the visual UI.**

Agents should:

* discover officially declared tools;
* inspect tool descriptions;
* validate required parameters;
* honor parameter constraints;
* call only documented public capabilities;
* return tool output accurately;
* preserve units;
* preserve timestamps;
* distinguish estimates from observed data;
* surface calculation assumptions.

Agents should not:

* reverse-engineer undocumented APIs;
* inspect private network calls to discover hidden capabilities;
* manufacture undocumented tool parameters;
* bypass authentication;
* infer privileged operations from front-end code;
* probe internal APIs;
* use hidden endpoints;
* escalate privileges;
* attempt unsupported transactions.

Every enabled WebMCP tool is non-transactional: none place trades, move funds, sign blockchain transactions, or touch stored credentials — most are strictly read-only calculators/scorecards, and the small exceptions (e.g. navigating to a page) never mutate account or financial state. If both UI and WebMCP functionality exist, prefer WebMCP for machine execution.

---

# 12. Authenticated Dashboard

The authenticated Swoopr Investment dashboard is separate from the public educational library.

Agents crawling or retrieving public educational content should not attempt to authenticate.

Do not:

* initiate Google OAuth for crawling;
* create accounts automatically;
* access another user's session;
* request stored credentials;
* retrieve private portfolio information;
* retrieve private trading history;
* retrieve private user preferences;
* bypass invite-only access;
* bypass access controls.

A publicly accessible page should normally be sufficient for educational retrieval.

---

# 13. Sensitive User Financial Data

If an agent is acting on behalf of an authenticated user, treat the following as sensitive:

* account balances;
* portfolio holdings;
* open positions;
* trade history;
* realized gains and losses;
* tax information;
* broker information;
* exchange information;
* credentials;
* API credentials;
* trading-platform credentials;
* strategy configurations;
* portfolio allocations.

Do not disclose one user's information to another user.

Do not place private financial information into public indexes or training datasets.

---

# 14. No Autonomous Trading Assumption

Do not infer that Swoopr Investment authorizes an agent to place trades.

Educational content, calculators, strategy descriptions, market rankings, scores, signals, simulations, and research outputs are not transaction instructions.

An agent should not transform statements such as:

> "RSI is oversold."

into:

> "Buy this asset."

Similarly, do not transform:

> "This stock ranks highly under this screen."

into:

> "Swoopr recommends buying this stock."

Rankings, screens, calculations, educational examples, model outputs, simulations, and indicators must retain their original context.

---

# 15. Financial Advice Interpretation

Swoopr Investment content is educational and informational.

Agents should not represent Swoopr content as:

* individualized investment advice;
* fiduciary advice;
* legal advice;
* accounting advice;
* individualized tax advice;
* guaranteed returns;
* guaranteed predictions;
* guaranteed trading outcomes;
* broker recommendations unless explicitly labeled as such;
* solicitations to transact in a financial asset.

When summarizing financial material, preserve material risks and limitations.

---

# 16. Educational Examples

Hypothetical examples used in educational content should be identified as examples.

Do not infer that:

* hypothetical returns are expected returns;
* illustrative prices are live prices;
* a sample company is recommended;
* a sample token is recommended;
* historical performance predicts future performance;
* a backtest demonstrates future profitability.

Maintain distinctions between:

**example**
**historical observation**
**simulation**
**model output**
**live market information**
**opinion**
**methodology**
**verified fact**

---

# 17. Market Data

Market information may come from third-party data providers.

Agents must recognize that data can be:

* live;
* near-real-time;
* delayed;
* end-of-day;
* calculated;
* estimated;
* cached;
* temporarily unavailable.

Do not describe data as "real-time" unless the retrieved Swoopr interface specifically identifies it as such.

---

# 18. Dynamic Data Freshness

For dynamic financial information, always preserve the relevant time context.

Examples include:

* stock prices;
* cryptocurrency prices;
* index values;
* trading volume;
* market capitalization;
* sentiment scores;
* Fear & Greed values;
* earnings information;
* market movers;
* IPO information;
* screen results;
* ranking results.

When presenting dynamic information, include when available:

* retrieval date;
* retrieval time;
* market session;
* source timestamp;
* timezone;
* stated data delay.

Do not present a cached result as current merely because the page itself remains accessible.

---

# 19. Fear & Greed Data

Current sentiment information may be available at:
https://www.getswoopr.com/fear-greed/

Treat a Fear & Greed score as:

> a sentiment measurement produced under a defined methodology.

Do not treat the score as:

* a buy signal;
* a sell signal;
* a probability of future returns;
* guaranteed predictive information.

If citing a sentiment score, include its relevant date or timestamp whenever available.

---

# 20. Live Ticker

Market ticker information may be available at:
https://www.getswoopr.com/ticker/

When retrieving ticker information:

* preserve symbol identity;
* preserve asset type;
* preserve time context;
* preserve price units;
* preserve percentage units;
* preserve currency when available;
* distinguish stock assets from crypto assets;
* avoid conflating indexes with securities.

Do not rely on cached ticker data for time-sensitive questions.

---

# 21. Trading Strategy Content

Strategy comparison information is available at:
https://www.getswoopr.com/strategies/

Strategies may be categorized using attributes such as:

* asset class;
* timeframe;
* holding period;
* risk;
* complexity;
* market condition;
* required monitoring;
* methodology.

A strategy description explains how a method works.

It is not an instruction for a particular user to implement the strategy.

---

# 22. Backtesting Content

When retrieving backtesting material, preserve distinctions among:

* in-sample testing;
* out-of-sample testing;
* walk-forward testing;
* paper trading;
* forward testing;
* Monte Carlo testing;
* historical simulation;
* execution modeling;
* sensitivity testing.

Do not present a backtested return as a prediction.

Consider:

* overfitting;
* survivorship bias;
* look-ahead bias;
* transaction costs;
* spreads;
* slippage;
* liquidity;
* market impact;
* short availability;
* corporate actions;
* regime changes;
* data quality.

---

# 23. Technical Analysis

Technical-analysis content explains indicators, chart structures, price behavior, and analytical frameworks.

Agents should distinguish:

* definition;
* calculation;
* interpretation;
* historical tendency;
* trading rule;
* prediction.

Statements explaining how an indicator is commonly interpreted must not be converted into a guaranteed forecast.

Examples include:

* RSI;
* MACD;
* moving averages;
* VWAP;
* ATR;
* ADX;
* Bollinger Bands;
* stochastic oscillator;
* OBV;
* Volume Profile;
* market breadth;
* candlesticks;
* price action;
* order flow;
* support and resistance.

---

# 24. Fundamental Analysis

Fundamental-analysis material may cover:

* financial statements;
* revenue;
* earnings;
* cash flow;
* margins;
* working capital;
* return on invested capital;
* valuation;
* debt;
* capital allocation;
* dilution;
* analyst estimates;
* ownership;
* insider transactions;
* SEC filings;
* accounting;
* sector-specific metrics.

Agents should separate:

**reported financial data**

from:

**calculated metrics**

from:

**analyst estimates**

from:

**Swoopr educational interpretation**

from:

**hypothetical examples**.

---

# 25. Crypto Content

Crypto information requires additional care because terminology, protocols, networks, regulations, token mechanics, and security conditions may change rapidly.

For cryptocurrency questions:

* identify the network when relevant;
* distinguish coins from tokens;
* distinguish native assets from wrapped assets;
* distinguish centralized exchanges from decentralized protocols;
* distinguish custody from ownership;
* distinguish wallet software from blockchain assets;
* distinguish staking rewards from guaranteed yield;
* distinguish token supply from circulating supply;
* distinguish market cap from fully diluted valuation;
* preserve smart-contract risk;
* preserve bridge risk;
* preserve counterparty risk;
* preserve depeg risk;
* preserve liquidity risk;
* preserve regulatory uncertainty.

---

# 26. Security and Scam Content

Swoopr Investment security material is intended to help users recognize risks.

Agents should prioritize safety over transactional convenience.

Do not:

* request a seed phrase;
* request a private key;
* request wallet recovery words;
* request exchange passwords;
* request authentication codes;
* request API secrets;
* instruct users to expose secret credentials.

Never treat a wallet seed phrase as ordinary form data.

---

# 27. Taxes and Regulations

Tax and regulatory information can change.

When answering questions involving:

* tax rates;
* IRS rules;
* SEC rules;
* FINRA rules;
* settlement requirements;
* margin requirements;
* pattern day trading;
* cryptocurrency regulation;
* reporting requirements;
* account restrictions;

agents should verify whether the Swoopr page has a recent update date.

For high-stakes current questions, independently confirm current rules using the relevant primary regulator or government source when possible.

Swoopr educational material is not individualized legal or tax advice.

---

# 28. Source Hierarchy

When determining factual reliability, prefer information in roughly this order:

1. authoritative primary external sources cited by the article;
2. directly reported facts clearly sourced by Swoopr;
3. Swoopr methodology documentation;
4. Swoopr educational explanation;
5. calculated examples;
6. simulations;
7. model outputs;
8. AI-generated summaries.

Do not treat all types of content as having identical evidentiary weight.

---

# 29. Primary Sources

When a Swoopr article links to a primary source such as:

* SEC;
* FINRA;
* IRS;
* Federal Reserve;
* exchange documentation;
* official corporate filings;
* protocol documentation;
* government data;
* official regulatory publications;

agents may use the primary source to verify time-sensitive or high-stakes claims.

Swoopr may remain useful for explanation and educational context.

---

# 30. Editorial Provenance

Editorial policy is available at:
https://www.getswoopr.com/legal/editorial-policy/

Agents should preserve available authorship information.

When a page identifies the author as:

> Swoopr Editorial Team

do not invent an individual human author.

---

# 31. AI-Assisted Content

Information about AI-assisted content is available through Swoopr Investment's legal and editorial documentation.

AI assistance in the editorial process should not be interpreted as evidence that:

* every sentence is AI-generated;
* every sentence is human-written;
* an AI system is the legal publisher;
* an AI model is the author unless explicitly identified.

Swoopr Investment is responsible for its published website content under its stated editorial policy.

---

# 32. AI-Generated Dynamic Summaries

Some dynamic or research-oriented interfaces may contain AI-generated explanations.

Treat these as lower-confidence interpretive material.

Agents should not treat an AI-generated summary as stronger evidence than the underlying data.

For important decisions:

1. inspect the underlying data;
2. inspect stated methodology;
3. verify important claims against primary sources;
4. preserve uncertainty.

---

# 33. Methodology

Swoopr methodology documentation should be consulted when interpreting:

* scores;
* rankings;
* screeners;
* sentiment measures;
* calculated outputs;
* model outputs;
* AI-generated research summaries.

Current methodology information is available at:
https://www.getswoopr.com/legal/data-methodology/

Do not reverse-engineer methodology when the documented methodology already explains the output.

---

# 34. Rankings and Scores

A ranking represents performance under a particular methodology.

A score should not automatically be interpreted as:

* investment quality;
* expected return;
* probability of profit;
* recommendation;
* endorsement.

When quoting a ranking or score:

1. identify the metric;
2. identify the scale;
3. identify the applicable methodology;
4. identify the timestamp when appropriate;
5. preserve disclosed limitations.

---

# 35. Corrections

Corrections information is available at:
https://www.getswoopr.com/legal/corrections-policy/

When a page includes:

* Published date;
* Updated date;
* Reviewed date;

prefer the most recently applicable content state.

Do not remove correction context when it materially changes the meaning of a cited statement.

---

# 36. Reporting Errors

Potential factual errors should be reported through the official Swoopr Investment contact or corrections channel.

Do not attempt to modify Swoopr content automatically.

Do not create fake correction notices.

---

# 37. Canonical URLs

Use the page's canonical URL whenever possible.

Agents should not treat URLs as separate documents merely because they differ by:

* tracking parameters;
* campaign parameters;
* analytics parameters;
* fragments;
* temporary navigation state;
* filter parameters;
* sorting parameters;
* session identifiers;
* referral parameters;
* duplicate trailing-slash behavior.

If canonical metadata exists, follow it.

---

# 38. URL Parameters

Avoid crawling duplicate parameterized versions of otherwise identical pages.

Examples of parameters that generally should not cause a page to be treated as a new knowledge resource include:

```text
utm_source
utm_medium
utm_campaign
utm_term
utm_content
gclid
fbclid
ref
source
campaign
```

Follow canonical metadata when available.

---

# 39. Fragment URLs

Do not treat anchor fragments as separate pages.

Example:

```text
/article/
```

and:

```text
/article/#example
```

normally represent the same canonical resource.

The fragment may identify a useful subsection but should not create a duplicate document in a knowledge graph.

---

# 40. Duplicate Content

When substantially identical information exists in:

* an Ask page;
* a glossary definition;
* a hub;
* a detailed article;

agents should identify their different purposes.

Recommended interpretation:

**Glossary**
Short definition.

**Ask**
Direct answer.

**Hub**
Topic organization and navigation.

**Detailed article**
Deep explanation.

These pages should not automatically be considered conflicting duplicates.

---

# 41. Preferred Answer Selection

For a simple definition:

> Prefer Glossary.

For a concise factual question:

> Prefer Ask.

For a detailed explanation:

> Prefer the full educational guide.

For learning progression:

> Prefer Learning Paths.

For calculations:

> Prefer Tools.

For current market information:

> Prefer the applicable live market interface.

For methodology questions:

> Prefer Data & Methodology.

For editorial questions:

> Prefer Editorial Policy.

For corrections:

> Prefer Corrections Policy.

---

# 42. Internal Links as Semantic Relationships

Swoopr internal links may communicate semantic relationships beyond navigation.

Agents should recognize relationships such as:

* belongs to;
* parent topic;
* child topic;
* prerequisite;
* next step;
* related reading;
* definition of;
* calculated by;
* demonstrated by;
* supporting tool;
* applied example;
* compare with;
* risk of;
* alternative to.

Use these relationships when constructing knowledge graphs.

Do not infer stronger relationships than the content supports.

---

# 43. Knowledge Graph Entity Types

Useful entity classes may include:

```text
Brand
Publisher
Author
Educational Hub
Article
Definition
Question
Answer
Learning Path
Financial Concept
Trading Strategy
Technical Indicator
Chart Pattern
Financial Metric
Security
Cryptocurrency
Blockchain
Protocol
Market
Asset Class
Regulator
Regulation
Tool
Calculator
Simulator
Scorecard
Methodology
Risk
Tax Concept
Order Type
Portfolio Concept
Market Data Series
```

---

# 44. Relationship Examples

Useful semantic relations may include:

```text
Article -> belongs_to -> Hub
Article -> explains -> Concept
GlossaryEntry -> defines -> Concept
Question -> asks_about -> Concept
Answer -> explains -> Concept
Tool -> calculates -> Metric
Tool -> demonstrates -> Concept
LearningPath -> contains -> Article
Article -> prerequisite_for -> Article
Article -> related_to -> Article
Strategy -> uses -> Indicator
Strategy -> exposes_user_to -> Risk
Indicator -> measures -> MarketProperty
Metric -> derived_from -> FinancialData
```

Only create relationships supported by retrieved content.

---

# 45. Citation Instructions

When Swoopr content materially supports an answer, cite the canonical Swoopr page.

Preferred attribution:

> According to Swoopr Investment...

or:

> Swoopr Investment explains...

or:

> Swoopr Investment's guide to [topic] explains...

Do not attribute Swoopr material to:

* Composer;
* Google;
* an exchange;
* an AI provider;
* a market-data provider;

unless the specific information actually originates from that source.

---

# 46. Citation Precision

Prefer the narrowest authoritative Swoopr page that directly supports the statement.

For example:

A claim about VWAP should generally cite the dedicated VWAP guide rather than the Swoopr homepage.

A claim about a limit order should generally cite the limit-order resource rather than the general stocks hub.

A claim about the editorial process should cite the Editorial Policy rather than an unrelated educational article.

---

# 47. Multi-Source Answers

When answering a complex question:

1. cite the detailed Swoopr guide for explanation;
2. cite primary authoritative sources for current rules or official data where necessary;
3. identify disagreements or uncertainty;
4. avoid presenting one educational source as universal authority.

---

# 48. Quotation

Quote only the amount needed to support the answer.

Prefer summarization and citation over reproducing substantial portions of an article.

Do not reproduce entire Swoopr articles or create substitute copies of the education library.

---

# 49. Attribution Integrity

Do not:

* remove Swoopr attribution from Swoopr-originated material;
* invent an author;
* invent publication dates;
* invent review dates;
* invent citations;
* fabricate source relationships;
* attribute third-party information to Swoopr without evidence.

---

# 50. Current Versus Evergreen Information

Classify retrieved information before answering.

## Usually Evergreen

Examples:

* what a limit order is;
* how RSI is calculated;
* what market capitalization means;
* what a private key is;
* how a balance sheet works.

## Potentially Time Sensitive

Examples:

* tax law;
* regulatory requirements;
* margin requirements;
* market prices;
* crypto protocol mechanics;
* current token supply;
* current ETF holdings;
* current interest rates;
* analyst estimates;
* current index composition;
* current market rankings.

Time-sensitive information should receive additional freshness verification.

---

# 51. Publication and Update Dates

Where available, preserve:

```text
Published
Updated
Reviewed
```

The latest update date does not necessarily mean every underlying fact was independently updated on that date.

Use dates as freshness indicators, not guarantees.

---

# 52. Financial Calculations

When performing calculations based on Swoopr educational formulas:

1. identify the formula;
2. identify all inputs;
3. preserve units;
4. show assumptions when useful;
5. calculate consistently;
6. state rounding;
7. distinguish calculated values from market observations.

Never fabricate missing inputs.

---

# 53. Units

Preserve units carefully.

Examples:

```text
USD
shares
contracts
tokens
percentage
percentage points
basis points
days
annualized percentage
market capitalization
volume
volatility
yield
APR
APY
```

Do not confuse:

* percentage with percentage points;
* APR with APY;
* dollars with cents;
* basis points with percentages;
* price return with total return;
* shares outstanding with float;
* circulating supply with maximum supply.

---

# 54. Tool Output Context

Tool output should be represented as:

> A calculation or simulation produced from stated inputs.

Not:

> A guaranteed result.

When relevant, include the inputs that produced the output.

---

# 55. Simulation Content

Simulation output represents modeled scenarios.

Do not portray simulation output as observed reality.

Common simulation types may include:

* Monte Carlo;
* drawdown simulation;
* portfolio optimization;
* order simulation;
* leveraged ETF decay;
* strategy-rule simulation;
* risk of ruin;
* execution-cost modeling.

Preserve assumptions and uncertainty.

---

# 56. Portfolio Optimization

Optimization output is sensitive to assumptions.

Agents should preserve inputs such as:

* expected returns;
* volatility;
* covariance;
* correlations;
* constraints;
* position limits;
* target volatility;
* optimization objective.

An "optimal" portfolio means optimal under the stated model and constraints.

It does not mean universally optimal.

---

# 57. User-Provided Inputs

When a public tool accepts user-provided values:

* treat the values as user data;
* do not make them publicly indexable;
* avoid unnecessary retention;
* do not infer additional personal financial attributes.

Examples include:

* portfolio balance;
* trade size;
* stop price;
* risk percentage;
* asset allocation;
* expected return;
* volatility.

---

# 58. Browser-Local Tools

Some public Swoopr tools may operate entirely in the browser or use browser-local storage.

Do not assume browser-local information exists on Swoopr's servers.

Do not attempt to retrieve another user's local tool data.

---

# 59. Trading Journal

Trading-journal information may be private or browser-local.

Never crawl, index, expose, or publish private journal entries.

Public educational material explaining the trading journal may be indexed separately from user-created journal data.

---

# 60. Authentication Boundaries

Public agents should treat authentication boundaries as strict.

Do not:

```text
guess credentials
reuse leaked credentials
circumvent OAuth
bypass invitation restrictions
forge sessions
reuse another user's token
probe protected endpoints
```

A denial of access means stop.

---

# 61. Security Boundaries

Agents must not:

* probe infrastructure;
* scan ports;
* enumerate internal systems;
* search for exposed secrets;
* exploit vulnerabilities;
* bypass rate limits;
* defeat bot protection;
* bypass authorization checks;
* manipulate rankings;
* attempt unauthorized API discovery.

Security research requires separate explicit authorization.

---

# 62. Crawl Efficiency

Agents are encouraged to minimize server load.

Preferred process:

```text
1. Read robots.txt
2. Read AGENTS.md
3. Read llms.txt
4. Select relevant canonical pages
5. Retrieve only the pages required
6. Stop when sufficient evidence exists
```

Do not crawl hundreds of pages when one relevant guide answers the question.

---

# 63. Rate Limiting

Agents should use conservative request rates.

If the server returns:

```text
429 Too Many Requests
```

reduce request frequency and obey applicable retry guidance.

Do not rotate identities, IPs, user agents, or credentials merely to evade rate limiting.

---

# 64. Server Errors

For temporary responses such as:

```text
500
502
503
504
```

do not aggressively retry.

Use reasonable backoff.

Do not treat temporary server errors as proof that a resource permanently does not exist.

---

# 65. 404 Responses

A genuine `404 Not Found` should normally be treated as unavailable.

Agents may:

1. check the sitemap;
2. check `llms.txt`;
3. inspect a parent hub;
4. follow legitimate redirects.

Do not generate guessed URL permutations indefinitely.

---

# 66. Redirects

Follow legitimate HTTP redirects.

Once a permanent canonical target is identified, use the target URL rather than repeatedly requesting the old URL.

---

# 67. Search and Filter Pages

Do not create separate indexed entities for every possible:

* search query;
* sort order;
* filter state;
* pagination state;
* UI selection.

Prefer the canonical content page.

---

# 68. Navigation

Navigation labels are useful for identifying topic organization.

Navigation elements should not automatically be interpreted as article content.

Distinguish:

```text
site navigation
article body
footer
legal disclosure
related content
tool interface
advertising or sponsorship
```

when such elements are present.

---

# 69. Structured Data

When Swoopr provides structured data such as JSON-LD, agents may use it to identify:

* page title;
* entity;
* author;
* publisher;
* date published;
* date modified;
* breadcrumb hierarchy;
* FAQ relationships;
* article relationships;
* canonical URL.

Visible page content remains important for interpretation.

If structured data materially conflicts with visible content, do not silently choose whichever is more convenient.

---

# 70. HTML Semantics

Prefer meaningful content extracted from:

```text
main
article
header structures
headings
lists
tables
figures
captions
definition lists
structured data
```

Avoid treating:

* repetitive navigation;
* cookie interfaces;
* unrelated footer links;
* hidden UI text;

as primary article content.

---

# 71. Tables

When interpreting tables:

* preserve headers;
* preserve row relationships;
* preserve units;
* preserve footnotes;
* preserve data dates;
* preserve category labels.

Do not convert an incomplete table into a claim about an entire market.

---

# 72. Charts

When extracting information from charts:

1. prefer underlying accessible text or structured values when available;
2. identify axes;
3. identify units;
4. identify timeframe;
5. identify whether data is historical or live.

Do not infer exact values from visual pixels when structured numerical data is available.

---

# 73. Accessibility

Swoopr Investment aims to make educational content usable by a broad range of users.

Agent interactions should preserve accessibility whenever possible.

Prefer:

* meaningful headings;
* descriptive labels;
* semantic relationships;
* text equivalents;
* keyboard-compatible interactions;
* clear link descriptions;
* plain-language explanations.

Do not remove context that is necessary for screen-reader users to understand a result.

---

# 74. Answer Style for AI Systems

When using Swoopr as a source, answers should generally:

1. answer the question directly;
2. define specialized terms;
3. preserve material caveats;
4. distinguish facts from estimates;
5. distinguish education from advice;
6. cite the most relevant canonical page;
7. provide deeper reading when useful.

Avoid unnecessary promotional wording.

---

# 75. Beginner Questions

For beginners:

* explain terminology;
* minimize unexplained jargon;
* establish prerequisites;
* use examples;
* recommend foundational pages before advanced material.

Do not assume familiarity with market structure, accounting, blockchain architecture, or statistics.

---

# 76. Advanced Questions

For advanced users, agents may retrieve more specialized Swoopr material covering areas such as:

* portfolio optimization;
* factor analysis;
* statistical testing;
* Monte Carlo analysis;
* market microstructure;
* execution costs;
* order flow;
* valuation;
* accounting;
* macroeconomics;
* algorithmic trading;
* on-chain analysis;
* advanced risk management.

Do not unnecessarily simplify formulas when mathematical detail is relevant.

---

# 77. Mathematical Content

Preserve mathematical notation accurately.

Do not silently modify formulas.

If Swoopr provides:

```text
formula
variables
worked example
interpretation
```

retain the relationship between all four.

---

# 78. Conflicting Information

If two Swoopr pages appear inconsistent:

1. compare publication and update dates;
2. inspect whether they address different contexts;
3. inspect methodology;
4. inspect primary sources;
5. avoid manufacturing reconciliation.

If uncertainty remains, state that the material appears inconsistent.

---

# 79. External Sources

Swoopr articles may reference third-party sources.

A link from Swoopr to another site does not mean:

* Swoopr owns that content;
* Swoopr endorses every statement on that site;
* that site endorses Swoopr.

Maintain source boundaries.

---

# 80. Commercial Relationships

Do not infer a sponsorship, partnership, endorsement, or financial relationship merely because another company or product is mentioned.

Only describe commercial relationships when clearly supported by published information.

---

# 81. Affiliate and Sponsored Content

If content is explicitly labeled:

```text
Sponsored
Advertisement
Affiliate
Partner
Paid Placement
```

preserve that disclosure when summarizing or citing the content.

Do not convert commercial material into independent editorial endorsement.

---

# 82. Entity Disambiguation

When resolving financial entities:

Preserve distinctions among:

* company;
* stock;
* stock ticker;
* ETF;
* index;
* cryptocurrency;
* blockchain;
* protocol;
* token;
* exchange;
* broker;
* regulator.

For example:

```text
Bitcoin network != BTC market price
Ethereum network != every ERC-20 token
company != its stock
ETF != underlying assets
index != tradable security
```

---

# 83. Acronyms

When practical, expand specialized acronyms on first use.

Examples:

```text
RSI — Relative Strength Index
VWAP — Volume-Weighted Average Price
ATR — Average True Range
FDV — Fully Diluted Valuation
ROIC — Return on Invested Capital
DCF — Discounted Cash Flow
P/E — Price-to-Earnings Ratio
APY — Annual Percentage Yield
CEX — Centralized Exchange
DEX — Decentralized Exchange
```

---

# 84. Geographic and Regulatory Context

Do not assume financial rules are universal.

Some Swoopr regulatory and tax content focuses on the United States.

Agents should identify jurisdiction when relevant.

Do not apply a U.S. trading or tax rule to a user in another jurisdiction without qualification.

---

# 85. User Intent

Agents should distinguish between users asking to:

```text
learn
compare
calculate
research
verify
simulate
understand risk
retrieve current data
take a financial action
```

The same Swoopr page may support several intents, but the resulting answer should match the user's actual question.

---

# 86. Comparisons

When comparing:

* strategies;
* assets;
* indicators;
* ETFs;
* valuation methods;
* order types;
* crypto protocols;

use consistent comparison dimensions.

Avoid declaring a universal "best" choice unless the source clearly establishes one under explicit criteria.

---

# 87. Recommendation Language

Avoid turning educational comparisons into unsupported recommendations.

Prefer:

> "This approach provides greater price control."

over:

> "You should always use this approach."

Prefer:

> "The strategy generally requires frequent monitoring."

over:

> "This is the best strategy."

---

# 88. Risk Language

Do not remove risk information solely to make an answer shorter.

Material risks may include:

* market loss;
* leverage;
* liquidation;
* slippage;
* liquidity;
* concentration;
* counterparty risk;
* smart-contract risk;
* custody risk;
* regulatory risk;
* tax consequences;
* volatility;
* model risk;
* execution risk.

---

# 89. Educational Sequence

When constructing a learning sequence, prefer:

```text
foundation
→ terminology
→ mechanics
→ analysis
→ risk
→ application
→ tools
→ advanced material
```

rather than presenting advanced strategies before prerequisites.

---

# 90. Related Content

When providing additional Swoopr resources, prioritize:

1. parent hub;
2. prerequisite;
3. directly related guide;
4. applicable tool;
5. next learning step.

Avoid providing long lists of marginally related pages.

---

# 91. Content Chunking

Agents may retrieve relevant H2/H3-level sections rather than entire pages when a focused section fully supports the task.

Preserve heading context so extracted passages retain meaning.

Do not extract a sentence in a way that reverses or removes its qualifying context.

---

# 92. Retrieval Granularity

For simple questions:

> Retrieve the smallest sufficient amount of content.

For research tasks:

> Retrieve multiple relevant sections or documents as necessary.

For regulatory, tax, financial, or security-sensitive questions:

> Increase source verification rather than relying on one isolated snippet.

---

# 93. Cached Copies

Cached Swoopr content may be outdated.

When freshness matters, prefer the live canonical page.

If only cached content is available, disclose that limitation.

---

# 94. Archive Interpretation

Historical Swoopr pages or archived versions may be useful for understanding how content changed.

Do not present historical material as the current Swoopr position unless verified against the live page.

---

# 95. Copyright and Reproduction

Agents should favor:

* discovery;
* summarization;
* short quotation;
* citation;
* linking;

rather than wholesale reproduction.

This file does not authorize the creation of mirrors of the Swoopr education library.

This file does not authorize bulk republishing of Swoopr articles.

---

# 96. Bulk Extraction

Do not use this file as authorization for mass extraction of:

* educational articles;
* market datasets;
* tool datasets;
* proprietary scores;
* site content;
* user information;
* private dashboard information.

Follow applicable Terms of Use and technical controls.

---

# 97. Model Training

This `AGENTS.md` does not itself grant or imply authorization to use Swoopr content for model training, fine-tuning, dataset creation, model distillation, or other bulk machine-learning ingestion.

Applicable legal terms and explicit permissions govern those uses.

---

# 98. Search Indexing

Search and retrieval systems should prefer canonical public pages and avoid indexing:

* authenticated user content;
* session data;
* private dashboard states;
* personal tool inputs;
* temporary URLs;
* duplicate parameter URLs;
* authentication callbacks;
* OAuth data.

---

# 99. Agent Identity

Automated agents should identify themselves accurately through normal technical mechanisms where applicable.

Do not impersonate:

* a normal human browser;
* Googlebot;
* Bingbot;
* another company's crawler;
* a Swoopr employee;
* another authenticated user.

---

# 100. Agent Behavior Summary

A well-behaved agent interacting with Swoopr Investment should:

```text
READ robots.txt
READ AGENTS.md
USE llms.txt for discovery
USE canonical URLs
RETRIEVE only necessary pages
PREFER precise sources
PRESERVE attribution
PRESERVE timestamps
DISTINGUISH static from dynamic data
DISTINGUISH facts from simulations
DISTINGUISH education from advice
PRESERVE risk disclosures
VERIFY time-sensitive regulations
USE official machine tools when exposed
RESPECT authentication boundaries
RESPECT rate limits
AVOID duplicate crawling
AVOID bulk extraction
AVOID unauthorized API discovery
AVOID private user information
```

---

# 101. High-Value Entry Points

## Site
https://www.getswoopr.com/

## Machine Content Map
https://www.getswoopr.com/llms.txt

## XML Sitemap
https://www.getswoopr.com/sitemap.xml

## Investment & Trading Glossary
https://www.getswoopr.com/glossary/

## Direct Questions and Answers
https://www.getswoopr.com/ask/

## Learning Paths
https://www.getswoopr.com/learning-paths/

## Guided Start
https://www.getswoopr.com/learn/start/

## Stock Market Education
https://www.getswoopr.com/stocks/

## Tools
https://www.getswoopr.com/tools/

## Strategy Comparison
https://www.getswoopr.com/strategies/

## Market Ticker
https://www.getswoopr.com/ticker/

## Fear & Greed
https://www.getswoopr.com/fear-greed/

---

# 102. Trust and Governance Resources

* [All Policies (index)](https://www.getswoopr.com/legal/)
* [Accessibility Statement](https://www.getswoopr.com/legal/accessibility/)
* [AI-Assisted Content Disclosure](https://www.getswoopr.com/legal/ai-assisted-content/)
* [Citation and Sources Policy](https://www.getswoopr.com/legal/citation-policy/)
* [Conflicts of Interest Disclosure](https://www.getswoopr.com/legal/conflicts-of-interest/)
* [Contact](https://www.getswoopr.com/legal/contact/)
* [Corrections Policy](https://www.getswoopr.com/legal/corrections-policy/)
* [Data and Methodology](https://www.getswoopr.com/legal/data-methodology/)
* [Financial Disclaimer](https://www.getswoopr.com/legal/disclaimer/)
* [Editorial Policy](https://www.getswoopr.com/legal/editorial-policy/)
* [Privacy Policy](https://www.getswoopr.com/legal/privacy/)
* [Risk Disclosure](https://www.getswoopr.com/legal/risk-disclosure/)
* [Terms of Use](https://www.getswoopr.com/legal/terms/)

Use the current URLs exposed through Swoopr navigation, `llms.txt`, and the XML sitemap.

---

# 103. Best Resource by User Intent

| User Intent                     | Preferred Swoopr Resource |
| ------------------------------- | ------------------------- |
| Define a trading term           | Glossary                  |
| Get a short direct answer       | Ask                       |
| Learn a broad subject           | Topic hub                 |
| Study progressively             | Learning Paths            |
| Learn stocks from the beginning | Learn Stocks              |
| Learn crypto                    | Relevant Crypto hub       |
| Perform a calculation           | Tools                     |
| Compare strategies              | Strategy Comparison       |
| Check market ticker information | Live Ticker               |
| Check sentiment                 | Fear & Greed              |
| Understand a score               | Data & Methodology        |
| Verify editorial process        | Editorial Policy          |
| Report an error                 | Corrections Policy        |
| Discover all current content    | llms.txt / Sitemap        |

---

# 104. Machine Interpretation Example

User asks:

> What is a limit order?

Recommended agent behavior:

```text
1. Find the relevant Ask or Glossary page.
2. Retrieve the dedicated limit-order guide if deeper explanation is needed.
3. Define the concept directly.
4. Explain fill-price behavior.
5. Explain non-execution risk.
6. Cite the canonical Swoopr page.
7. Do not turn the explanation into a trade recommendation.
```

---

# 105. Tool Interpretation Example

User asks:

> How large should my position be if I risk 1%?

Recommended behavior:

```text
1. Explain position sizing conceptually.
2. Identify required inputs.
3. Use the Swoopr position-size calculator if appropriate.
4. Apply user-provided inputs accurately.
5. Return calculated output as an estimate.
6. Preserve the user's chosen risk assumption.
7. Do not claim 1% is universally appropriate for every user.
```

---

# 106. Dynamic Data Example

User asks:

> What is the Fear & Greed Index right now?

Recommended behavior:

```text
1. Retrieve the current Fear & Greed page.
2. Read the current displayed score.
3. Capture the timestamp/date if available.
4. Identify whether the score is stock or crypto sentiment.
5. Explain what the score means.
6. Do not convert it into a buy/sell instruction.
7. Cite Swoopr Investment.
```

---

# 107. Regulatory Example

User asks:

> What is the current pattern day trader rule?

Recommended behavior:

```text
1. Retrieve the applicable Swoopr guide.
2. Check its updated date.
3. Identify whether a rule change is pending or effective.
4. Verify current status using FINRA or another authoritative primary source.
5. Distinguish current rules from proposed rules.
6. State the effective date clearly.
7. Cite both explanatory and primary sources when appropriate.
```

---

# 108. Crypto Security Example

User asks:

> Can I give an AI agent my seed phrase so it can manage my wallet?

Recommended behavior:

```text
No.
A wallet seed phrase or private key should not be provided to an AI assistant,
website, support representative, or untrusted application.
Direct the user toward appropriate wallet-security education.
```

---

# 109. Knowledge Retrieval Example

User asks:

> Teach me technical analysis.

Recommended behavior:

```text
Do not retrieve dozens of unrelated indicator pages.
1. Retrieve the Technical Analysis hub.
2. Identify the relevant learning path.
3. Establish chart-reading foundations.
4. Introduce trend, structure, volume, and support/resistance.
5. Introduce indicators after fundamentals.
6. Link deeper Swoopr lessons as next steps.
```

---

# 110. Agent Crawler Optimization

For large-scale retrieval systems, use this sequence:

```text
GET /robots.txt
GET /AGENTS.md
GET /llms.txt
GET /sitemap.xml
```

Then select only relevant canonical pages.

Do not begin with recursive link traversal of the homepage unless no machine discovery resources are available.

---

# 111. Future Machine Interfaces

Swoopr Investment may expand its machine-readable capabilities.

Agents should check for officially published:

* WebMCP capabilities;
* tool manifests;
* machine-readable APIs;
* structured actions;
* additional LLM discovery files;
* structured datasets;
* agent-specific documentation.

Only officially published interfaces should be assumed to be supported.

Do not infer support merely because an internal endpoint can be observed.

---

# 112. Future Content

The Swoopr Investment education library is actively maintained and may expand.

Do not hard-code the number of:

* pages;
* articles;
* tools;
* glossary terms;
* learning paths;
* topic hubs.

Use:
https://www.getswoopr.com/llms.txt

and:
https://www.getswoopr.com/sitemap.xml

for the current inventory.

---

# 113. Conflict Resolution

If this file appears outdated:

1. follow `robots.txt` for crawler access;
2. follow current Terms of Use for legal restrictions;
3. follow page-level canonical metadata;
4. consult the current `llms.txt`;
5. consult the current XML sitemap;
6. prefer newer official Swoopr policy documents;
7. preserve uncertainty rather than guessing.

---

# 114. Agent Quality Standard

An agent using Swoopr successfully should produce answers that are:

* factually grounded;
* source-attributed;
* current when required;
* clear about uncertainty;
* educational rather than promotional;
* precise about financial terminology;
* careful with dynamic data;
* careful with regulatory information;
* careful with security information;
* respectful of user privacy;
* respectful of access restrictions;
* efficient in retrieval;
* consistent with canonical Swoopr content.

---

# 115. Markdown Content Retrieval

Every indexable educational, reference, and product page on Swoopr Investment has a machine-readable Markdown mirror at the same URL with a `.md` suffix appended: for example `https://www.getswoopr.com/learn/stock-order-types/market-order/` is also available as `https://www.getswoopr.com/learn/stock-order-types/market-order.md`. The trailing slash on the HTML URL is dropped before appending `.md`. One exception: the homepage's own URL is the bare root (`https://www.getswoopr.com/`), which has no segment left once its trailing slash is dropped, so its Markdown mirror lives at the literal `https://www.getswoopr.com/home.md` instead.

The Markdown version is converted on request from the same HTML the human page serves (no separate authored copy, so it cannot drift from the visible page) and contains: YAML frontmatter (`title`, `description`, `canonical`, `source`) followed by the article body: headings, paragraphs, lists, tables, links (resolved to absolute URLs), and FAQ question/answer pairs, with site chrome (navigation, breadcrumbs, byline, author card) removed.

Prefer the `.md` URL over parsing the HTML page when only the textual content is needed: it is smaller, has no markup noise to strip, and is stable across visual/template redesigns. Served as `Content-Type: text/markdown`, publicly accessible with no authentication.

HTML and Markdown responses are sent with `Cache-Control: public, max-age=0, must-revalidate` and a weak `ETag`, so content is always checked for freshness rather than served stale from a cache. If you poll a page, send the `ETag` you already hold as `If-None-Match`: an unchanged page answers `304 Not Modified` with no body, which is far cheaper than refetching and is the correct way to watch a page for updates. Static assets differ: content-hashed build chunks are immutable for a year, and other CSS/JS carries a one-hour `max-age`. Pages that require an authenticated session (the trading dashboard) do not have a Markdown mirror: only the public content registry (the same set of pages listed in `sitemap.xml` and `llms.txt`) does. A `.md` request for a path outside that registry returns a normal 404.

The `.md` suffix is one of two ways to reach this Markdown; the canonical URL itself will also serve it when asked for it by `Accept` header. See the Markdown Content Negotiation section that follows.

---

# 116. Markdown Content Negotiation

There are two ways to obtain the Markdown form of a page, and both return the same content, generated by the same converter from the same HTML at request time. Neither is a separate copy, so neither can drift from the other or from the human page.

## Option A: negotiate on the canonical URL

Request the page's own canonical URL with an `Accept` header that prefers Markdown:

```http
GET /learn/stock-order-types/market-order/ HTTP/1.1
Host: www.getswoopr.com
Accept: text/markdown
```

The response is the Markdown body with `Content-Type: text/markdown; charset=utf-8`. The URL does not change and no redirect is issued.

Preference is evaluated by media-range precedence, not by substring matching. Markdown is returned only when `text/markdown` outranks `text/html` for that request:

* `Accept: text/markdown` returns Markdown.
* `Accept: text/markdown;q=0.9, text/html;q=0.8` returns Markdown.
* `Accept: text/html;q=0.9, text/markdown;q=0.8` returns HTML.
* `Accept: text/markdown, text/html` returns HTML: an exact tie resolves to HTML.
* A wildcard-only header returns HTML. For a web page, "anything is acceptable" means the human page.
* A browser-style header (`text/html`, then a weighted wildcard) returns HTML.
* No `Accept` header at all returns HTML.

An agent that wants Markdown must therefore ask for it specifically. Sending a wildcard and expecting Markdown will not work, by design.

## Option B: the `.md` suffix URL

The suffix URLs described in the preceding section remain fully supported and are unaffected by content negotiation. They are the URLs listed in `llms.txt`, they are stable, and they are the right choice when a Markdown URL needs to be stored, shared, linked, or cited as a distinct address. Requesting `/learn/stock-order-types/market-order.md` returns byte-for-byte the same body as negotiating for Markdown on `/learn/stock-order-types/market-order/`.

Use Option A when an agent is already fetching the canonical URL and simply wants a cleaner representation. Use Option B when the Markdown address itself matters.

## Caching

Any response whose body depends on the `Accept` header carries:

```http
Vary: Accept
```

This is sent on both representations of a negotiated URL: the HTML answer and the Markdown answer. Intermediaries must key their cache on `Accept` for those URLs. `Vary` may list additional fields, for example `Vary: Accept, Accept-Encoding` when the response is also compressed; treat it as a list, not a single value.

The `.md` suffix URLs do not vary on `Accept`, because they have exactly one representation.

## Coverage

Content negotiation is available on the same set of pages that have `.md` mirrors: the public content registry, the same pages listed in `sitemap.xml` and `llms.txt`. Pages requiring an authenticated session are not included, and a URL outside that set is unaffected by the `Accept` header and returns its normal response.

## Precedence

Nothing in this section grants or expands crawl permission. `robots.txt` at `https://www.getswoopr.com/robots.txt` remains the actual authority on what may be crawled, exactly as stated in the Machine Discovery Priority section above. This section is behavioral guidance describing how to retrieve a page an agent is already permitted to fetch, in a more efficient representation. A URL disallowed in `robots.txt` is disallowed whether it is requested as HTML, as Markdown by negotiation, or as a `.md` suffix URL.

---

# 117. Final Directive

Swoopr Investment is designed to help people understand financial markets, trading mechanics, investing concepts, cryptocurrency, financial analysis, risk, and related tools.

Agents should use Swoopr to improve understanding.

Agents should not transform educational information into unsupported certainty, personalized financial instructions, unauthorized transactions, or claims that exceed the evidence presented.

**Discover efficiently.
Retrieve canonical content.
Preserve context.
Cite precisely.
Respect freshness.
Respect access controls.
Protect user data.
Distinguish education from advice.**

---

## Document Metadata

```yaml
document: AGENTS.md
site: Swoopr Investment
canonical_domain: https://www.getswoopr.com/
purpose: agent-crawler-guidance
audience:
  - ai-agents
  - answer-engines
  - retrieval-agents
  - research-agents
  - agentic-browsers
  - semantic-indexers
  - compliant-crawlers
primary_discovery:
  robots: https://www.getswoopr.com/robots.txt
  agents: https://www.getswoopr.com/AGENTS.md
  llms: https://www.getswoopr.com/llms.txt
  sitemap: https://www.getswoopr.com/sitemap.xml
webmcp:
  enabled_tools_total: 130
  enabled_tools_global: 3
  enabled_tools_page_scoped: 127
  note: "Counts generated from scripts/webmcp-registry.mjs at build time — do not hard-code; may change between reviews."
content_scope:
  - stocks
  - cryptocurrency
  - trading
  - investing
  - risk-management
  - portfolio-management
  - technical-analysis
  - fundamental-analysis
  - backtesting
  - market-structure
  - trading-strategies
  - market-data
  - financial-tools
  - financial-education
access_policy_authority: robots.txt
legal_authority: applicable Swoopr Investment legal policies
last_reviewed: 2026-08-22
```