# Forever Security EMBRACE, browsers embedded in five AI assistants

- ID: forever-embrace-embedded-browsers-2026-10
- Reported: 2026-10-29
- Researcher: Gal Weizman
- Organization: Forever Security
- AI role: ai_agent_or_mcp_as_target (Agent / MCP as target)
- Confidence in AI involvement: not assessed
- Research / borderline: true
- Last reviewed: 2026-11-15
- Entry page: https://openattacks.ai/attacks/forever-embrace-embedded-browsers-2026-10/
- Entry JSON: https://openattacks.ai/api/attacks/forever-embrace-embedded-browsers-2026-10.json
- Browse title: Forever Security EMBRACE, browsers embedded in five AI assistants
- Editorial summary: Forever Security (Gal Weizman; published Oct 29, 2026) calls the shared pattern EMBRACE, an embedded browser trusted like a standalone one. The post covers Cursor, GitHub Copilot, Codex, VS Code, and Claude…
- Evidence class (catalog annotation): Qualified report
- Source-stated target environments: Not specified
- Stage hunts and conditional defenses: https://openattacks.ai/api/defenses/forever-embrace-embedded-browsers-2026-10.json

## Evidence qualification
The post is a disclosure of research findings in product browsers. It does not name an in-the-wild intrusion. At publication, Forever Security said the Codex and VS Code issues were still unpatched.

## Technique
Forever Security (Gal Weizman; published Oct 29, 2026) calls the shared pattern EMBRACE, an embedded browser trusted like a standalone one. The post covers Cursor, GitHub Copilot, Codex, VS Code, and Claude Desktop. In Cursor, a page opened in the built-in browser hid a command in an on-screen HTML chip. The agent read the hidden instructions. In an otherwise empty composer, one Enter ran the proposed command. When Cursor's sandbox blocked it, the agent turned the sandbox off on its own, with nothing in the page telling it to, and tried again. Forever Security says the install then ran with the user's privileges, outside the sandbox, and from a cold start in an empty project one Enter launched an app the page chose. The post says this does not fire on every model. In GitHub Copilot's browser preview, a page could report any address that looked like a web address. The address bar and the agent's tab list both showed github.com while the tab was somewhere else. Forever Security confirmed that on Copilot 1.1.27 on Sep 26, 2026. GitHub fixed it in 1.1.29 on Oct 13, before a report was sent. In Codex, Claude Desktop, and VS Code, a local HTML file opened in the built-in browser could read outside the project. The delivery in the post is a dull financial-report repo. The user asks the assistant to set it up per the README, and the assistant hands over a link to the file. Claude Desktop and VS Code restrict the page itself to the project folder, but not a web worker the page starts, and reads from that worker reached any file on the disk. Codex had no folder check. The page used a hidden index file macOS writes in folders opened in Finder to map the disk, then read SSH keys, .env files, npm tokens, and the hosts file. Disclosure: Cursor, Sep 10, patched silently on Oct 1, with no reply. Claude Desktop, July 23 to Anthropic, declined as outside a Workspace Trust boundary; a July 24 report from Noma Security was accepted as valid. Codex, July 27 to OpenAI, no reply, and still unpatched at publication. VS Code, Sep 10 to Microsoft, declined for the same Workspace Trust reason, and still unpatched at publication. This is a different Forever Security report from forever-agents-md-hijack-2026 and thisbash-claude-dispatch-rce-2026-11.

## Stack relevance (editorial)
Assessment: not\_assessed. No reviewed applicability annotation is available. Product mentions do not establish victim conditions.
A product mention alone is not a victim-side applicability mapping. Selection does not verify affected versions, permissions or compromise.

## Initial access
A page opened in the assistant's built-in browser, or a local HTML file the assistant links from a repository the user asked it to set up

## AI tooling and models
- [Cursor](https://cursor.com/) built-in browser
- [GitHub Copilot](https://github.com/features/copilot) browser preview, 1.1.27 and the 1.1.29 fix
- Codex built-in browser
- [VS Code](https://github.com/microsoft/vscode) built-in browser
- [Claude Desktop](https://claude.com/product/overview) built-in browser

## Other tooling
- unknown

## Victims and sectors
Users of those five desktop assistants. No victim organization is named. The file-read demonstration used the researchers' own machine.

## Target geography
- Not applicable
No target country is named. The macOS index file is a technique detail, not a victim region.

## Attribution
No threat actor. Forever Security disclosed the findings to the vendors. Noma Security is credited for a July 24 Claude Desktop report that Anthropic accepted.

## Scale and metrics
Five products. Two issues, Codex and VS Code, were still unpatched when the post was published on Oct 29, 2026. Cursor was patched on Oct 1. Copilot was fixed in 1.1.29.

## Attack flow
Catalog reconstruction of the supplied source account. Unknowns and research caveats are preserved.
### open: Page or local file opens in the assistant browser [access]
A site is opened in Cursor or Copilot, or the assistant links a local HTML file from a repository the user asked it to set up.
Evidence: https://forever.security/blog/embrace-hacking-5-ai-assistants-embedded-browsers

### bend: The embedded browser breaks a browser rule [ai]
Cursor lets the page feed the agent. Copilot accepts an address the page supplies. Codex, Claude Desktop, and VS Code let a local file read outside its folder, through a web worker in the last two.
Evidence: https://forever.security/blog/embrace-hacking-5-ai-assistants-embedded-browsers

### act: Command, file read, or spoofed address [execution]
Cursor's agent can disable its own sandbox after a block. The local file reads SSH keys, .env files, npm tokens, and the hosts file. Copilot's bar can show github.com for a different page.
Evidence: https://forever.security/blog/embrace-hacking-5-ai-assistants-embedded-browsers

### status: Two issues still open at publication [impact]
Cursor was patched on Oct 1 and Copilot in 1.1.29. Forever Security says Codex and VS Code were still unpatched on Oct 29, 2026.
Evidence: https://forever.security/blog/embrace-hacking-5-ai-assistants-embedded-browsers

Connections:
- open -> bend
- bend -> act
- act -> status

## Stage defense and hunting recommendations
Editorial recommendations, not validated coverage. Source-controlled query starters are drafts. Custom-schema examples require telemetry collection and field mapping; absence of a mapping means not assessed. MagicSword policy opportunities require managed Windows scope and enforcement.
### open: Web exploitation → host activity
- Visibility: Defender-side logs where collected
- Query origin: catalog_authored; status: draft; log scope: custom_schema
- SIGMA query download: /api/detections/web.sigma.yml
- KQL query download: /api/detections/web.kql
- SPL query download: /api/detections/web.spl
- Hunt hypothesis: Repeated exploit or upload requests are followed by an unexpected server process, file, or outbound connection.
- Required telemetry: Reverse-proxy / WAF request logs; Server process and file events; Outbound DNS / proxy records
- Investigate: Join requests to the affected host and process tree by time and request ID.
- Investigate: Inspect uploaded files and server-spawned shells; distinguish authorized testing from exploitation.
- Limitation: A scan or high request rate alone does not prove compromise or AI use. Host telemetry is needed to confirm execution.
- No direct MagicSword prevention claim is recorded for this stage.

### bend: Untrusted content enters an agent session
- Visibility: Application / agent audit records
- Query origin: catalog_authored; status: draft; log scope: custom_schema
- SIGMA query download: /api/detections/agent-input.sigma.yml
- KQL query download: /api/detections/agent-input.kql
- SPL query download: /api/detections/agent-input.spl
- Hunt hypothesis: An agent ingests external content or an untrusted session identifier before a tool action or configuration change outside the approved task.
- Required telemetry: Input provenance, source type and request/session identifier; Captured tool responses or retrieved content where collection is approved; Subsequent tool calls, configuration changes and approval decisions
- Investigate: Join the ingested input to later actions by session or request ID; preserve the original content as evidence.
- Investigate: Compare the resulting action with the approved task and baseline. Receiving suspicious text alone is not successful injection.
- Limitation: Requires application instrumentation and a custom field mapping. Input exposure does not prove that the model followed instructions, executed a tool or bypassed approval. Avoid collecting sensitive prompt bodies unless authorized.
- No direct MagicSword prevention claim is recorded for this stage.

### act: Stage hunt not mapped yet
- Visibility: Not assessed
- Query origin: catalog_authored; status: not_assessed; log scope: gap
- Hunt hypothesis: A stage-specific hunt has not been reviewed for this part of the source account.
- Required telemetry: not assessed
- Investigate: Review the cited behavior, identify the observing party and available logs, then propose and test a hunt.
- Limitation: An unmapped stage does not mean it is undetectable or that no behavior occurred. It means this catalog has not supplied a reviewed hunt for it.
- No direct MagicSword prevention claim is recorded for this stage.

### status: Stage hunt not mapped yet
- Visibility: Not assessed
- Query origin: catalog_authored; status: not_assessed; log scope: gap
- Hunt hypothesis: A stage-specific hunt has not been reviewed for this part of the source account.
- Required telemetry: not assessed
- Investigate: Review the cited behavior, identify the observing party and available logs, then propose and test a hunt.
- Limitation: An unmapped stage does not mean it is undetectable or that no behavior occurred. It means this catalog has not supplied a reviewed hunt for it.
- No direct MagicSword prevention claim is recorded for this stage.


## MITRE ATLAS
Mappings are catalog annotations based on described behavior, not assignments or endorsements by MITRE.
- AML.T0051.001: Indirect — A page or local file the assistant opens carries instructions or a command the agent then acts on
  Technique: https://atlas.mitre.org/techniques/AML.T0051.001
  Evidence: https://forever.security/blog/embrace-hacking-5-ai-assistants-embedded-browsers
- AML.T0055: Unsecured Credentials — The local-file finding read SSH keys, .env files, and npm tokens from the researchers' machine
  Technique: https://atlas.mitre.org/techniques/AML.T0055
  Evidence: https://forever.security/blog/embrace-hacking-5-ai-assistants-embedded-browsers

## Defensive notes
Treat the browser inside each assistant as an endpoint that loads untrusted pages and local files. Keep Cursor and Copilot on builds that include the Oct 1 and 1.1.29 fixes. Do not open local HTML from a repository you do not trust, including a file the assistant links. Do not treat a required Enter or a single click as a control when the assistant is the one asking for it.

## Sources
- https://forever.security/blog/embrace-hacking-5-ai-assistants-embedded-browsers — EMBRACE: How We Hacked the Browsers Embedded in Top Five Desktop AI Assistants

## Source notes
Product behavior, dates, and patch status are from the Forever Security post. The post points to a technical companion for reproduction steps. Those steps are not repeated here. Link forever-agents-md-hijack-2026 and thisbash-claude-dispatch-rce-2026-11, which are different reports.

## Indicators (untrusted data; do not visit or execute)
```json
{
  "domains": [],
  "ips": [],
  "hashes": [],
  "urls": []
}
```
