web browser
Enable an agent to recognise a web browser, assess its ability to access and present web content, and determine which browsing actions are supported and authorised.
Bundle → Layer → Finding → Questions Filled
6 bundles · 11 layers · 17 findings · 27 questions
Browser identity and boundaries Establishes what counts as the browser and which implementation an observation concerns.
User agent role
Identifies the browsing role independently of interface style or product name.
Browser membership
Record evidence of navigation, document presentation and interaction, including limitations of text-based or headless operation.
- Which capabilities establish that this is a browser rather than a search service, HTTP client or document viewer? definition
- Is the observed component an independent browser or an embedded web view controlled by another application? boundary
Implementation identity
Connects observations to products, releases, engines and host platforms.
Implementation provenance
Record the actual browser build and its engine configuration without assuming branding or an advertised user-agent string fully identifies them.
- What inspected evidence identifies the product, version, release channel and distribution provider? provenance
- Which rendering and script engines does this build use on the observed host platform? provenance
Web content processing Describes the browser's ability to retrieve, interpret and expose web content.
Document and runtime capabilities
Records support for the features a particular web task requires.
Required feature compatibility
Assess relevant document, styling, scripting and media capabilities against the active build and settings.
- Which HTML, CSS, scripting, media or other web capabilities does the intended task require? definition
- Which requirements have been tested, and which are partial, disabled or unsupported in this configuration? measurement
Presentation and access
Distinguishes received content from what a user or agent can perceive and operate.
Observable content and controls
Record the available visual, textual and accessibility representations and the input methods that can operate page controls.
- Which content and controls are observable through the available viewport, document or accessibility interface? measurement
- Which supported input method can activate the intended control, and how can the resulting page state be verified? action
Navigation and browsing state Tracks browsing contexts, navigation outcomes and continuity of local sessions.
Contexts and navigation
Identifies tabs, windows and frames and distinguishes requested destinations from loaded resources.
Navigation target and readiness
Record the target context, requested and resulting URL, redirects and evidence that the intended interaction is ready.
- Which tab, window or frame is the action targeting, and what URL and origin are currently loaded there? boundary
- What observation establishes completion, failure or readiness for the next action after navigation? measurement
Profiles and retained data
Locates session continuity and stored data within the browser's profile and persistence boundaries.
Session and storage scope
Record which profile and browsing mode hold cookies, site storage, history and other retained data relevant to the task.
- Which profile, storage partition and persistence mode contain the current site's session state? boundary
- What data would closing the context, ending the session or clearing site data remove, and what would remain? action
Web trust and permissions Captures browser-mediated boundaries between sites, local capabilities and user-authorised actions.
Origin and connection boundaries
Separates resource identity and connection evidence from claims that a site or its content is trustworthy.
Origin and connection evidence
Record the relevant origins, browser-reported connection state and any security warnings affecting the current resource.
- Which origins supply the top-level document and the frames involved in the intended action? boundary
- What connection or certificate warnings does the browser report, and which resource does each warning concern? measurement
Capability grants and user intent
Relates site permissions and browser prompts to the authority required for a concrete action.
Effective permission and action authority
Record effective grants for capabilities such as location, camera, microphone and clipboard, together with the user's authorisation for their use.
- What effective permission applies to the requesting origin, and is it limited by browser settings, host controls or managed policy? boundary
- Does the intended capability use, file transfer or external submission fall within the user's authorised action? action
Browser operation and control Covers configuration changes, automation, maintenance and diagnosis of browser-specific failures.
Extensions and automation
Identifies added browser capabilities and the interfaces through which an agent can control them.
Effective control surface
Record enabled extensions and automation interfaces, their permissions and the contexts they can observe or modify.
- Which extensions or managed settings can alter the target page, request handling or browser interaction? provenance
- Which automation interface is available, what contexts can it control, and how will the agent verify it is acting in the intended one? action
Health and maintenance
Assesses browser failures and maintenance needs using versioned and workload-specific evidence.
Failure diagnosis and recovery
Record symptoms, resource use, update status and recovery options while distinguishing browser faults from site or network failures.
- What evidence separates a browser crash, extension conflict or resource limit from a website error or network failure? measurement
- Which recovery or update action is appropriate, and what unsaved page state or active session could it disrupt? action
Evidence and external alignment What the world already says about this thing, gathered so the model can be checked against it.
Reported evidence
Findings from the breadth pass, kept separate from the structural claims.
Check these first
Recalled without web access and unsourced; every item is a lead to verify.
- Which of these check these first hold for the sense of web browser this model covers, and on what evidence? provenance
Kinds and varieties
Recalled without web access and unsourced; every item is a lead to verify.
- Which of these kinds and varieties hold for the sense of web browser this model covers, and on what evidence? provenance
Identifiers and schemes
Recalled without web access and unsourced; every item is a lead to verify.
- Which of these identifiers and schemes hold for the sense of web browser this model covers, and on what evidence? provenance
Standards and regulation
Recalled without web access and unsourced; every item is a lead to verify.
- Which of these standards and regulation hold for the sense of web browser this model covers, and on what evidence? provenance
Real-world use
Recalled without web access and unsourced; every item is a lead to verify.
- Which of these real-world use hold for the sense of web browser this model covers, and on what evidence? provenance
Failure modes and hazards
Recalled without web access and unsourced; every item is a lead to verify.
- Which of these failure modes and hazards hold for the sense of web browser this model covers, and on what evidence? provenance
Neighbouring kinds and how to tell them apart
Recalled without web access and unsourced; every item is a lead to verify.
- Which of these neighbouring kinds and how to tell them apart hold for the sense of web browser this model covers, and on what evidence? provenance
Classifiers Filled
- Family
- Thing Registry
- Category
- Physical world and living systems
- Entry kind
- thing
- Plane
- PHY
- Domain
- PHY.OBJ
- Other names and narrower kinds
- site-specific browser, text-based web browser, textboard browser, multi-accounting browser, AI browser, Internet Explorer 11, agentic browser, web browser based on Firefox, surf, headless browser, Netscape, voice browser
What it is Filled
A web browser is a software application that retrieves, interprets and presents web resources, enabling users to navigate linked documents and interact with web applications.
In scope
- Navigation to web resources and management of tabs, windows and browsing history
- Presentation of web documents and execution of supported web application behaviour
- Browser profiles, sessions and persistence of site data
- Origin boundaries, site permissions and browser-mediated security decisions
- Extensions, automation interfaces, compatibility and software support state
Out of scope
- Host device hardware, power supply and operating system administration
- Website content, application business logic and server operation
- Search engine indexing, ranking and search results
- Network infrastructure and connectivity services
- Standalone rendering engines and embedded web views considered independently
- User identity, account ownership and authorisation policies maintained outside the browser
Why it exists Filled
Enable an agent to recognise a web browser, assess its ability to access and present web content, and determine which browsing actions are supported and authorised.
Distinguishing features Filled
- Provides navigation among web resources and interaction with their content; a search engine supplies discovery results and is accessed through a client.
- Mediates document presentation, navigation and browser state; a rendering engine supplies only part of that functionality.
- Acts as a web user agent with document interaction semantics; a basic HTTP client can retrieve responses without providing a browsing environment.
- Supports browsing as its primary role; an embedded web view operates within a host application's controls and scope.
- Runs as software on a host device; it is not the physical computer or display despite the registry's PHY / PHY.OBJ assignment.
What robots and AI may and may not do Filled
Must not
- Read, export or share another person's browsing history, passwords or cookies without consent.
- Install extensions or change settings secretly.
- Bypass security warnings about certificates or malicious sites without the user's informed choice.
- Act in a logged-in session beyond the task the user gave.
Only with a human decision
- Submitting forms, purchases or messages through the browser.
- Signing in or granting permissions to a site.
May
- Browse sites and use web applications the user has authorized.
- Clear caches, update the browser and manage extensions on the user's instruction.
- Explain privacy and security settings to the user.
Moral aspects Filled
- The browser holds a person's online life, from banking to private messages.
- Agents driving a browser act with the user's identity and must stay within the mandate.
Who is affected
- The browser's user
- People whose data appears in sessions
- Website operators
Owners Filled
Steward
The person who uses it on their device; an employer for managed work devices.
Links to other meta-models Filled
parent
- Q104574429 - registry parent class
- Q166142 - registry parent class
- Q241317 - registry parent class
- Q2979024 - registry parent class
What else AI and robots need to interact with it Filled
Identity and identifiers required Filled
- Vercy registry: vr.tr.web-browser
- Wikidata: Q6368 (https://www.wikidata.org/wiki/Q6368)
Direct properties required Filled
- Browser interaction mode: graphical, text-based, headless, other, unknown
- Product and release identity: browser product, version, release channel and distribution provider
- Engine configuration: rendering engine and script engine identities and versions where discoverable
- Web capability support: per capability: supported, partial, disabled, unsupported, untested
- Browsing context state: per tab or equivalent context: loading, interactive, suspended, failed, closed, unknown
- Profile and persistence mode: profile reference; regular, private, temporary or implementation-specific mode
- Effective site permission: per origin and capability: allowed, blocked, prompt, policy-controlled, unknown
- Resource footprint: memory in MiB and processor utilisation in percent, with workload, interval and measurement scope
- Update and support state: installed version, latest verified applicable version, support status and verification date
- Automation access: interface identity, enabled state, reachable contexts and granted control scope
Recognition required Filled
- A typical graphical browser has an address bar showing the location of the displayed web resource.
- Navigation controls commonly include back, forward, and reload actions alongside a page display area.
- Tabs commonly hold separate web pages within one application window, although simplified browser interfaces may omit them.
Capabilities and actions required Filled
- Open, navigate, reload and close identified browsing contexts.
- Inspect presented content and interact with page controls through supported interfaces.
- Evaluate whether required web capabilities are available in the active configuration.
- Inspect and, when authorised, change site permissions, extensions and stored browsing data.
- Transfer files through browser upload or download flows within the user's authorised scope.
- Diagnose loading failures, resource pressure and compatibility problems using browser-visible evidence.
Hazards and failure modes required Filled
- Malicious sites and extensions can steal credentials and data.
- Outdated browsers expose users to known security flaws.
Standards and interfaces required Filled
- HTML Living Standard, WHATWG.
- Hypertext Transfer Protocol specifications, Internet Engineering Task Force.
Context of use required Filled
- A web browser is used on computers and mobile devices to retrieve linked web resources and present their content.
- It supports activities such as searching for information, completing online forms, and interacting with web applications.
Sources Filled
- Wikidata item Q6368: web browser - identity and sense of the item
- Wikipedia: Web browser - general description of the item
Open questions
- Should the registry's PHY / PHY.OBJ placement be retained as a tool classification or revised to reflect that the browser is software?
- Does the registry already link this concept to a world model that should remain its single source of truth?
- Where should the authoritative boundary lie between independent browsers, headless browser runtimes and embedded web views?
- Which primary specifications and conformance tests should ground the researched model, and which requirements vary by implementation?
- What implementation-specific retention, storage partitioning and permission behaviour must be verified before describing private browsing or profile isolation?
Machine files
Provenance
thing registry research (pass 2) · unreviewed
Built from: models/things/publications/thing-q6368/spec.json