High School IntermediateModule I8Lesson 5 of 8

I8.5 Request Forgery, Clickjacking, and User Action Protection

Learn how fictional applications protect sensitive browser actions through anti-forgery tokens, SameSite cookies, origin checks, correct methods, clear confirmations, reauthentication, frame restrictions, idempotency, cross-origin review, evidence-based investigation, validation, and professional closure.

Lesson Progress

Request Forgery, Clickjacking, and User Action Protection

High School IntermediateI8: Web Security Defense • Lesson 5 of 8

63% complete

Readiness Check

Before You Start

0/5 ready

Professional Hook

A Valid Session Does Not Prove the User Intended the Action

A fictional browser can hold a valid finance session while another page attempts to submit an invoice approval. A hidden frame can place a sensitive button under a harmless-looking control. A duplicate request can create two payments unless transaction state is designed carefully. Defenders protect the connection between identity, browser context, user intent, authorization, and final business state.

Weak design

“The fictional browser sent a valid session cookie, so accept the state-changing request immediately.”

Strong design

“Verify token, origin, method, authorization, user confirmation, reauthentication, transaction state, and final business result.”

Objective 1

Explain how fictional authenticated browser state, cookies, origins, forms, links, frames, and user actions can combine to create request-forgery or clickjacking risk.

Objective 2

Distinguish cross-site request forgery, clickjacking, open redirects, unsafe state-changing GET requests, confused-deputy behavior, and ordinary user error.

Objective 3

Identify defensive controls such as anti-forgery tokens, SameSite cookies, origin and referer checks, reauthentication, confirmations, frame protections, and transaction review.

Objective 4

Evaluate fictional browser, application, session, policy, user, and business evidence without creating harmful demonstrations or testing real sites.

Objective 5

Create a professional fictional User Action Protection Review with findings, owners, remediation, validation, monitoring, and residual risk.

Why This Matters

Sensitive Actions Need Proof of Context and Intent

Fictional payment approval, account recovery, data sharing, enrollment, permission change, export, and deletion can be high impact even when they use a valid account and session. Strong systems verify the request came through the approved application, show the user exactly what will happen, require stronger proof when needed, prevent duplicates, record the final state, and allow safe review.

Threat Patterns

Eight Ways User Actions Can Be Misrepresented

Cross-site form submission

A fictional user is signed in to the payroll portal while another page attempts to submit a change request to it.

Risk

The browser may include session cookies even though the user did not intentionally use the payroll portal.

Strong defense

Anti-forgery token, SameSite cookie policy, origin checks, server-side authorization, and confirmation for sensitive changes.

Evidence

Origin, cookie context, token result, route, session, user action, application decision, and system-of-record state.

State-changing GET request

A fictional delete or approval action is triggered by following a link or loading an image-style URL.

Risk

Browsers, crawlers, previews, and users may activate the action unintentionally.

Strong defense

Use appropriate non-GET methods, require anti-forgery validation, authorization, confirmation, and idempotent processing.

Evidence

Method, route, request source, user session, action result, and database state.

Hidden framed button

A fictional sensitive button is visually covered by another page element while the protected application is embedded underneath.

Risk

The user may click an action they cannot clearly see or understand.

Strong defense

Frame-ancestors or equivalent protection, clear confirmation, reauthentication, and user-visible context.

Evidence

Frame policy, parent origin, browser behavior, click coordinates, action route, and final result.

Transparent overlay

A fictional page places an invisible or nearly invisible interface over a harmless-looking control.

Risk

The user’s click can be redirected to a hidden application action.

Strong defense

Restrict framing, avoid sensitive single-click actions, require confirmation, and monitor unexpected framed use.

Evidence

Browser DOM, frame tree, policy result, user report, action request, and application logs.

Open redirect in trusted application

A fictional trusted portal accepts an arbitrary destination and sends users to an unapproved site.

Risk

The trusted domain can be used to make deceptive navigation look legitimate.

Strong defense

Server-defined route names or exact destination allowlists with clear user confirmation.

Evidence

Received destination, normalized host, allowlist decision, redirect response, browser navigation, and owner approval.

Repeated or duplicate submission

A fictional payment or enrollment form is submitted twice because of retry behavior or repeated clicks.

Risk

The same intended action may create duplicate business state.

Strong defense

Idempotency key, server-side duplicate detection, transaction state, confirmation, and visible completion status.

Evidence

Request IDs, idempotency key, user, amount or object, transaction IDs, and final system state.

Cross-origin API request

A fictional browser page from one origin attempts to call a protected API at another origin.

Risk

Overbroad cross-origin configuration may expose sensitive responses or actions to unintended browser contexts.

Strong defense

Exact origin allowlists, narrow methods and headers, credential restrictions, anti-forgery controls, and authorization.

Evidence

Origin, preflight, allowed origin, credentials mode, method, route, response, and application result.

Misleading approval flow

A fictional user sees a generic Continue button without clear recipient, amount, permission, or effect.

Risk

The user may approve an action without understanding the actual business consequence.

Strong defense

Explicit confirmation, trusted context, reauthentication, separation of duties, and transaction review.

Evidence

Displayed confirmation, target, amount or scope, user action, approval chain, and system-of-record result.

User Action Protection

Eight Controls That Work Together

Anti-forgery token

Require a fictional server-validated value tied to approved application state before accepting sensitive requests.

Benefit

Helps distinguish requests created by the legitimate application flow from unrelated cross-site submissions.

Limitation

Tokens must be unpredictable, validated correctly, scoped appropriately, and protected from unsafe rendering or leakage.

Validation

Confirm approved requests succeed and requests with missing, invalid, reused, or mismatched tokens are denied safely.

SameSite cookie policy

Limit when a fictional browser sends session cookies from cross-site contexts.

Benefit

Reduces unintended authenticated requests from unrelated sites.

Limitation

Required identity, payment, and integration flows may need carefully designed exceptions.

Validation

Test approved same-site and cross-site workflows in supported browsers without weakening sensitive routes.

Origin and referer validation

Compare the fictional request source with approved application origins for sensitive operations.

Benefit

Adds server-side context about where the browser request came from.

Limitation

Headers can be absent or affected by browser, proxy, and privacy behavior, so fallback design is required.

Validation

Confirm approved origins work, unapproved origins are denied, and missing-header cases follow documented policy.

Correct HTTP methods

Use fictional GET for retrieval and non-GET methods for state-changing actions.

Benefit

Reduces accidental activation by links, crawlers, previews, and browser loading behavior.

Limitation

Method choice alone does not provide anti-forgery or authorization protection.

Validation

Confirm state-changing GET requests are rejected and approved methods still require all other controls.

Sensitive-action confirmation

Show the fictional user the exact action, target, amount, recipient, permission, or data effect before completion.

Benefit

Reduces hidden, misleading, accidental, and duplicate actions.

Limitation

Generic confirmations can become meaningless if they do not show trusted details.

Validation

Test that users can identify the real action and that canceling causes no state change.

Reauthentication or step-up verification

Require fresh fictional identity proof before selected high-risk actions.

Benefit

Limits the value of an unattended or older session.

Limitation

It should be risk based and accessible without creating unnecessary friction or unsafe bypasses.

Validation

Confirm normal low-risk actions remain usable and high-risk actions require fresh approved verification.

Frame restrictions

Limit whether fictional pages may be embedded by other origins.

Benefit

Reduces clickjacking and deceptive overlay opportunities.

Limitation

Approved embedded workflows require exact, owned, and tested exceptions.

Validation

Confirm sensitive pages cannot be framed by unapproved origins while required approved embeds still work.

Idempotency and transaction state

Prevent fictional duplicate requests from creating repeated actions.

Benefit

Protects payments, enrollments, approvals, and other high-impact workflows from retries and double clicks.

Limitation

Keys, time windows, object scope, and final-state handling must be designed carefully.

Validation

Submit safe duplicate test requests and confirm one intended business result.

Core Concept

Separate Six Questions

Session

Which fictional account and browser session were attached to the request?

Source

Which fictional origin, frame, page, or application created the request?

Request

Which fictional method, route, token, target, and values were sent?

Intent

Did the fictional user clearly see and approve the exact action?

Decision

Which fictional authorization, token, origin, confirmation, and transaction controls acted?

Impact

Which fictional database, payment, permission, account, or business state actually changed?

Cross-Origin Review

Eight Areas to Review Before Allowing Browser Access

Allowed origins

Strong pattern

Exact fictional application origins are listed and owned.

Weak pattern

Any origin is reflected or a broad wildcard is used with sensitive access.

Evidence

Origin, configuration, response header, application owner, and browser result.

Review question

Which exact browser applications need cross-origin access?

Credentials

Strong pattern

Fictional credentialed requests are allowed only for approved origins and routes.

Weak pattern

Credentials are allowed broadly without a precise origin and business need.

Evidence

Credentials mode, cookie behavior, allowed-origin response, route, and session context.

Review question

Does the cross-origin request require user session state at all?

Methods

Strong pattern

Only required fictional methods are allowed.

Weak pattern

All methods are permitted by default.

Evidence

Preflight request, allowed methods, route implementation, and application result.

Review question

Which operations are actually required by the approved client?

Headers

Strong pattern

Only required fictional request headers are accepted.

Weak pattern

Any header is allowed without review.

Evidence

Requested headers, allowed headers, application parser, and security controls.

Review question

Could custom headers influence identity, tenant, role, or routing decisions?

Readable response

Strong pattern

Only approved fictional response data is exposed to the calling origin.

Weak pattern

Sensitive data is returned because the server assumes browser policy provides authorization.

Evidence

Response fields, account, route, origin, authorization, and browser access result.

Review question

Would the same data be allowed if the request came from the approved application directly?

Preflight handling

Strong pattern

Fictional preflight checks are consistent, narrow, cached safely, and tied to actual route behavior.

Weak pattern

The preflight allows more than the real application intends.

Evidence

OPTIONS request, response headers, cache duration, origin, method, and later request.

Review question

Does the preflight accurately describe the permitted action?

Error handling

Strong pattern

Denied fictional origins receive safe responses without sensitive detail.

Weak pattern

Errors reveal internal origin lists, routes, credentials, or stack details.

Evidence

Response status, body category, internal log, request ID, and redaction result.

Review question

Can defenders investigate the denial without exposing internal details to the client?

Monitoring

Strong pattern

Fictional cross-origin denials, unexpected origins, policy changes, and sensitive-route use are monitored.

Weak pattern

No owner reviews broad origin changes or repeated denied requests.

Evidence

Change record, event schema, dashboard, alert, owner, and review result.

Review question

Who owns the cross-origin policy and validates changes?

Evidence Matrix

What Request and User-Action Evidence Can Prove

Evidence source

Browser request record

Can support

The fictional origin, method, route, cookie context, headers, frame context, and user interaction.

Limitation

Browser evidence may be incomplete and does not alone prove server acceptance or business impact.

Evidence source

Application anti-forgery log

Can support

The fictional token presence, match result, session binding, route, decision, and rejection reason.

Limitation

A token pass does not replace authorization, confirmation, or transaction controls.

Evidence source

Session and cookie metadata

Can support

The fictional cookie attributes, issue time, scope, same-site behavior, session, account, and expiration.

Limitation

Cookie metadata does not prove the user intended the request.

Evidence source

Frame and content-policy record

Can support

The fictional parent origin, embedding decision, frame restriction, browser violation, and page.

Limitation

A frame-policy event does not by itself prove a user clicked or a protected action completed.

Evidence source

Application authorization record

Can support

The fictional account, role, resource, route, action, policy, result, and reason.

Limitation

Authorization success does not prove clear user understanding or business approval.

Evidence source

Transaction or database record

Can support

The fictional object, amount, recipient, approval, transaction ID, commit, rollback, and final state.

Limitation

The data layer may not explain how the browser request was initiated.

Evidence source

User and owner report

Can support

The fictional visible page, intended action, understood confirmation, expected workflow, and business consequence.

Limitation

Human memory and interpretation require technical correlation.

Evidence source

Code and configuration review

Can support

The fictional token logic, origin policy, cookie settings, frame restrictions, method handling, confirmation, and idempotency design.

Limitation

Review should be confirmed against deployed behavior and current application versions.

Defensive Workflow

Protect Sensitive Browser Actions in Six Steps

1

Identify sensitive actions

List fictional create, update, delete, approve, pay, share, reset, enroll, export, and permission-changing operations.

2

Map browser state

Review fictional cookies, sessions, origins, frames, methods, forms, redirects, and cross-origin behavior.

3

Review request protections

Evaluate fictional anti-forgery tokens, SameSite settings, origin checks, authorization, and safe method design.

4

Protect user intent

Require fictional clear confirmation, trusted context, reauthentication, separation of duties, and idempotency where appropriate.

5

Correlate impact

Compare fictional browser, application, session, transaction, user, and business records before claiming completion or harm.

6

Remediate and validate

Apply narrow fictional fixes, test approved and denied flows, monitor, define rollback, and document residual risk.

Case Classification

Eight Outcomes with Different Evidence Requirements

Expected protected action

The fictional request uses the approved application flow, session, method, token, origin, authorization, confirmation, and transaction controls.

Required documentation

Record user, session, route, token result, origin, action, target, approval, transaction, and owner.

Cross-site request denied

The fictional browser or application rejects an unintended cross-site state-changing request.

Required documentation

Record origin, cookie context, token, route, decision, user state, and confirmation of no business change.

Request-forgery weakness

The fictional application accepts a sensitive request without sufficient anti-forgery, origin, method, confirmation, or authorization controls.

Required documentation

Record affected routes, sessions, actions, scope, owners, containment, remediation, and validation.

Clickjacking protection working

The fictional browser or application prevents sensitive pages from being framed by unapproved origins.

Required documentation

Record frame policy, parent origin, page, browser result, approved exceptions, and tests.

Clickjacking weakness

A fictional sensitive action can be visually disguised or overlaid because framing and confirmation controls are insufficient.

Required documentation

Record page, frame tree, user-visible context, action route, affected scope, and safe remediation.

Duplicate action prevented

Fictional idempotency or transaction state stops repeated requests from creating multiple business results.

Required documentation

Record request IDs, idempotency key, transaction IDs, final state, and validation.

Legitimate exception

A fictional cross-origin or embedded workflow is verified, narrowly scoped, owned, tested, monitored, and required.

Required documentation

Record exact origins, routes, methods, credentials, owners, expiration, validation, and rollback.

Evidence incomplete

The fictional browser, frame, session, token, transaction, user, or business evidence is insufficient for a reliable conclusion.

Required documentation

State the gap, confidence, temporary control, owner, due date, and decision criteria.

Correlated Request Timeline

Follow a Fictional Invoice Approval Attempt from Browser to Closure

15:10:01

Session

A fictional finance user is signed in to the approved invoice portal with a valid secure session.

Authenticated browser state exists before the later request.

15:12:14

External page

The user opens a fictional supplier article in another tab.

A separate browser origin becomes active.

15:12:16

Cross-site request

The external page attempts to submit a hidden invoice-approval request to the finance portal.

A request-forgery pattern is attempted.

15:12:17

Cookie policy

The fictional session cookie is withheld because the request is cross-site and the cookie uses a restrictive SameSite policy.

One browser-layer control reduces authenticated request context.

15:12:18

Application

The finance portal also rejects the request because the anti-forgery token is missing and the origin is unapproved.

Server-side protections independently deny the action.

15:12:19

Transaction

No fictional invoice approval, payment state, or database transaction is created.

No confirmed business impact follows the denied request.

15:13:02

Monitoring

A structured event records the session, route, unapproved origin, missing token, denial, and request ID.

Defenders receive useful evidence without storing sensitive content.

15:18:00

Review

The team finds a legacy invoice route that still accepts state-changing GET requests.

A separate control weakness exists even though the first request was denied.

15:22:00

Containment

The legacy route is disabled and requests are redirected to a safe review page.

Narrow containment prevents accidental activation.

15:29:00

Remediation

The application replaces the route with POST, anti-forgery validation, authorization, explicit invoice confirmation, and idempotency.

Method, browser, application, user-intent, and duplicate-action controls are combined.

15:36:00

Frame review

Sensitive finance pages receive restrictive frame-ancestor protection.

Clickjacking defense is added to the same workflow.

15:44:00

Positive test

An approved finance user reviews the exact invoice and completes one approval after reauthentication.

Legitimate sensitive use remains available.

15:48:00

Negative test

Missing-token, wrong-origin, framed, GET, duplicate, and canceled requests create no approval.

Several unsafe conditions are denied.

Day 7

Monitoring

No new cross-site approval attempts succeed, and normal invoice processing remains stable.

Short-term validation supports closure.

Key Vocabulary

Request and User-Action Protection Terms

Cross-site request forgery

A defensive category in which a fictional authenticated browser is caused to send an unintended state-changing request to another application.

Clickjacking

A defensive category in which a fictional user is visually misled into clicking or interacting with a hidden or disguised interface.

Anti-forgery token

A fictional unpredictable value tied to approved application state and validated by the server before sensitive requests are accepted.

Origin

A fictional combination of scheme, host, and port used by browsers and servers to describe where a request or page comes from.

Referer

A fictional request header that may identify the previous page or source location, subject to browser and privacy behavior.

SameSite cookie

A fictional cookie attribute that limits when browsers send the cookie from cross-site contexts.

State-changing request

A fictional request that creates, updates, deletes, approves, pays, shares, resets, enrolls, or otherwise changes application or business state.

Reauthentication

A fictional requirement to prove identity again before a sensitive action or after elevated risk.

Confirmation step

A fictional user-facing review that clearly states the action, target, amount, recipient, or permission before completion.

Frame protection

A fictional browser and server control that restricts whether a page may be embedded inside another page.

Confused deputy

A fictional condition where a trusted application performs an action using its authority after receiving an untrusted or misleading request.

Idempotency

A fictional design property that prevents duplicate submissions from creating repeated unintended business actions.

Fake Dashboard

Fake User Action Protection Dashboard

Training dashboard for the fictional Meadowbrook finance and enrollment applications.

Sensitive routes

37

Fictional payment, approval, enrollment, reset, export, sharing, deletion, and permission routes.

Denied cross-site requests

84

Missing-token, wrong-origin, framed, unsafe-method, expired-session, and canceled-action cases.

Open findings

5

Legacy GET action, broad origin exception, weak confirmation, duplicate-action, and evidence-gap reviews.

Fake SOC Alert

Cross-Site Invoice Approval Attempt Denied

Source: Fake User Action Protection Console • Time: 03:12 PM

High Severity
A fictional finance user has a valid session when an unrelated browser origin attempts to submit a hidden invoice-approval request. The session cookie is withheld by a restrictive SameSite policy, and the application independently rejects the request because the anti-forgery token is missing and the origin is unapproved. No approval or payment transaction is created.
Defensive recommendation: Preserve the browser, session, token, origin, application, and transaction evidence; review every sensitive route; remove state-changing GET behavior; require anti-forgery validation, authorization, clear confirmation, reauthentication, frame protection, and idempotency; then test approved and denied workflows.

Fake Log Panel

Fake Request-Forgery and Remediation Timeline

training-log-viewer.log
15:10:01 SESSION account='finance-104' app='invoice-portal' state='valid'
15:12:14 BROWSER external_origin='supplier-article.example'
15:12:16 CROSS_SITE_REQUEST route='/invoice/approve' method='POST'
15:12:17 COOKIE same_site='restrictive' sent='false'
15:12:18 APPLICATION token='missing' origin='unapproved' result='deny'
15:12:19 TRANSACTION approval='none' payment='none' database_change='none'
15:13:02 MONITOR event='structured' sensitive_content='not_stored'
15:18:00 REVIEW legacy_route='/invoice/approve-link' method='GET'
15:22:00 CONTAIN legacy_route='disabled'
15:29:00 REMEDIATE method='POST' token='required' confirmation='explicit' idempotency='enabled'
15:36:00 FRAME_POLICY finance_pages='restricted'
15:44:00 POSITIVE_TEST approved_user='pass' reauth='pass' approval_count='1'
15:48:00 NEGATIVE_TEST missing_token='deny' wrong_origin='deny' framed='deny' duplicate='one_result'
DAY7 MONITOR successful_cross_site_approvals='0' normal_processing='stable'

Training note: this is fake data for defensive analysis practice only.

Analyze the Evidence

Which Request-Forgery Conclusion Is Best Supported?

The fictional finance user has a valid session.
An unrelated browser origin attempts to submit an invoice approval.
The restrictive SameSite policy prevents the session cookie from being sent.
The application also rejects the request because the anti-forgery token is missing.
The origin is not on the approved application list.
No approval, payment, or database transaction is created.
A separate legacy GET route is discovered and disabled.
After remediation, approved finance use succeeds while missing-token, wrong-origin, framed, GET, canceled, and duplicate requests do not create extra approvals.

Which conclusion is strongest?

Common Mistakes

Mistakes That Weaken User Action Protection

Using fictional GET requests for actions that create, update, delete, approve, pay, share, or change permissions.
Relying on SameSite cookies as the only request-forgery defense.
Validating anti-forgery tokens without also enforcing authentication, authorization, and business rules.
Accepting anti-forgery values from predictable or unrelated session state.
Trusting origin or referer headers without a documented missing-header policy.
Allowing broad cross-origin access with credentials.
Treating a content-policy or frame denial as proof that a user clicked or a transaction completed.
Hiding a sensitive action instead of providing clear confirmation and reauthentication.
Failing to use idempotency for payments, enrollments, approvals, and retries.
Creating broad frame or origin exceptions for an entire provider when only one owned application needs access.
Closing after a denied request without confirming no database, payment, permission, or business state changed.
Building harmful demonstrations or testing real sites instead of using safe fictional evidence.

Safe Practice Lab

Complete a Fictional User Action Protection Review

Fictional Evidence Set

Meadowbrook Sensitive Action Review

Review forty supplied fictional records covering sessions, cookies, origins, forms, tokens, methods, frames, cross-origin policy, confirmations, reauthentication, authorization, idempotency, transactions, user reports, remediation, validation, monitoring, and closure.

Required Analysis

  1. Inventory every fictional state-changing route and business effect.
  2. Map session, cookie, origin, method, token, frame, confirmation, and authorization controls.
  3. Review cross-origin origins, credentials, methods, headers, responses, errors, and monitoring.
  4. Separate attempted request, user intent, server decision, transaction state, and business impact.
  5. Identify unsafe GET, missing-token, broad-origin, weak-confirmation, framing, and duplicate-action cases.
  6. Write findings with facts, confidence, alternatives, owners, remediation, positive tests, negative tests, rollback, and residual risk.
Use only supplied fictional evidence. Do not create hidden forms, overlays, framing demonstrations, cross-site requests, or duplicate transactions against real applications. Do not access sessions, cookies, tokens, accounts, payments, permissions, or private data.

Scenario Decision Lab

A Sensitive Action Uses GET

A fictional legacy application approves a request when a user follows a specially formatted link. The route changes business state and does not require an anti-forgery token or confirmation.

Scenario Decision Lab

An Approved Support Widget Needs Cross-Origin Access

A fictional support widget from one owned origin needs to read a limited help-status API from another owned origin. The current configuration allows any origin and credentials.

Defender Habits

Request Forgery, Clickjacking, and User Action Protection Checklist

Check Your Understanding

I8.5 Mini Quiz: Request Forgery, Clickjacking, and User Action Protection

Choose your answers first. Explanations appear only after submission.

1. What is the core risk in fictional cross-site request forgery?

2. Which defense combination is strongest for a fictional sensitive form?

3. Why should state-changing actions avoid GET?

4. What is the primary purpose of frame restrictions?

5. Why is idempotency useful for fictional payments and approvals?

6. A fictional cross-site request is denied because the token is missing. What additional evidence is needed before closing?

7. Which cross-origin configuration is strongest?

Portfolio Prompt

Portfolio Prompt

Create a fictional User Action Protection Review using at least forty session, cookie, origin, method, token, frame, cross-origin, confirmation, reauthentication, authorization, idempotency, transaction, user, business, remediation, validation, monitoring, and closure records. Include a sensitive-route inventory, request-protection matrix, frame and cross-origin review, normalized timeline, findings, owners, positive tests, negative tests, rollback, evidence gaps, residual risk, and closure criteria.

Use only fictional applications, routes, accounts, sessions, origins, tokens, transactions, users, and organizations.
Include one denied cross-site request, one state-changing GET weakness, one framing weakness, one duplicate-action case, and one approved cross-origin exception.
Keep valid session, request source, user intent, server decision, transaction state, and business impact separate.
Do not create or run hidden forms, overlays, frames, cross-site requests, or duplicate transactions.

Key Takeaways

What You Should Remember

1.A fictional valid session does not prove the user intended a sensitive action.
2.Anti-forgery tokens, SameSite cookies, origin checks, authorization, confirmation, and reauthentication solve different parts of the problem.
3.State-changing GET routes can be activated by links, crawlers, previews, and normal browser behavior.
4.Frame restrictions and clear confirmations reduce clickjacking and hidden-action risk.
5.Idempotency protects payments, approvals, enrollments, and retries from duplicate results.
6.Strong closure validates approved and denied flows, transaction state, legitimate use, monitoring, evidence gaps, residual risk, and owner approval.

Navigation

Continue Module I8