mirror of
https://github.com/openclaw/openclaw.git
synced 2026-06-03 20:14:06 +00:00
7920af0c9e
* feat(browser): add optional vision understanding to screenshot tool
* fix(browser): wrap vision output as external content, enforce maxBytes, forward auth profiles
* fix(browser): remove no-op scope/attachments config, drop profile pass-through lacking runtime support
* feat(media-understanding): add profile/preferredProfile to DescribeImageFileWithModelParams and forward to describeImage
* style(browser): add curly braces to satisfy eslint curly rule
* fix(browser): correct tools.browser.enabled help text to match actual behavior
* fix(browser): thread agentDir/workspaceDir from plugin tool context into browser vision
* refactor(browser): move vision config from tools.browser to browser.models
The browser plugin's vision configuration now lives on the top-level
`browser` config namespace (browser.models, browser.visionEnabled,
browser.visionPrompt, etc.) instead of `tools.browser`. This aligns
with the plugin's existing config location and avoids confusion between
tool-level and plugin-level settings.
- Remove tools.browser from ToolsSchema and ToolsConfig
- Add models/vision* fields to BrowserConfig and its zod schema
- Update getBrowserVisionConfig to read from cfg.browser
- Update schema help, labels, and quality test
- Update vision.test.ts to use new config shape
* docs(browser): add screenshot vision configuration section
Document the new browser.models config for automatic screenshot
description via vision models, enabling text-only main models to
reason about web page content.
* fix(browser): remove deliverable media markers from vision result, drop unused import
P1: Vision-success path no longer exposes the raw screenshot as
deliverable media (removes MEDIA: line and details.media.mediaUrl).
This prevents channel delivery from auto-sending sensitive page content
when the intended output is a text description.
P2: Remove unused ToolsMediaUnderstandingSchema import that would fail
noUnusedLocals typecheck.
* fix(browser): add command/args fields to browser models schema
The browser vision model schema uses .strict(), so CLI-type entries
with command/args were rejected by TypeScript. Add these fields to
align with MediaUnderstandingModelSchema.
* chore(browser): remove debug console.log statements
* fix(browser): harden screenshot vision result against MEDIA: directive injection and restore image sanitization on failure fallback
ClawSweeper #84247 review round 2:
P1 (security, high): neutralize line-start MEDIA: directives in vision descriptions
before wrapping with wrapExternalContent. The agent media extractor scans every
browser tool-result text block via splitMediaFromOutput which treats line-start
MEDIA: as a trusted local-media delivery directive, and browser is on the
trusted-media allowlist. Without neutralization, page or vision-provider output
containing 'MEDIA:/tmp/secret.png' could synthesize a channel-deliverable media
artifact from untrusted content. wrapExternalContent itself does not strip
line-start directives. Introduce neutralizeMediaDirectives in vision.ts that
prepends '[neutralized] ' to any line whose trimStart() begins with MEDIA:
(case-insensitive), defanging the parser anchor while keeping the original
text human-readable.
P2 (compatibility): pass resolveRuntimeImageSanitization() to imageResultFromFile
in the vision-failure catch fallback. The non-vision screenshot path already
forwards this option (d5cc0d53b7) so configured agents.defaults.imageMaxDimensionPx
takes effect. Without this fix, any provider timeout/error silently bypasses the
sanitization guard and returns a raw full-resolution screenshot.
Regression coverage:
- vision.test.ts: 6 unit cases for neutralizeMediaDirectives (no-op fast path,
mid-line MEDIA: untouched, line-start defanged, leading-whitespace defanged,
case-insensitive, multiple directives per blob).
- browser-tool.test.ts: 2 integration cases that drive the full screenshot
tool execute path:
- 'neutralizes MEDIA: directives in vision text and does not attach media'
asserts no line matches /^\s*MEDIA:/i in returned text, secret path text
is preserved verbatim, details.media is absent, and imageResultFromFile
is not called on the success path.
- 'preserves screenshot image sanitization on vision failure fallback'
mocks describeImageFileWithModel to reject and asserts the fallback
imageResultFromFile call receives imageSanitization: {maxDimensionPx:1600}
plus the 'browser screenshot vision failed' extraText.
* fix(browser): apply clawsweeper fallback media fix from PR #84247
* refactor: reuse media image understanding for browser screenshots
* refactor: use structured media delivery
* test: update music completion media instruction expectation
* fix: trim buffered reply directive padding
* test: refresh codex prompt snapshots for message media aliases
---------
Co-authored-by: scotthuang <scotthuang@tencent.com>
Co-authored-by: Peter Steinberger <steipete@gmail.com>
3.1 KiB
3.1 KiB
summary, read_when, title
| summary | read_when | title | ||
|---|---|---|---|---|
| Rich output protocol for structured media, embeds, audio hints, and replies |
|
Rich output protocol |
Assistant output can carry a small set of delivery/render directives:
- structured
mediaUrl/mediaUrlsfields for attachment delivery [[audio_as_voice]]for audio presentation hints[[reply_to_current]]/[[reply_to:<id>]]for reply metadata[embed ...]for Control UI rich rendering
Remote media attachments must be public https: URLs. Plain http:,
loopback, link-local, private, and internal hostnames are ignored as attachment
directives; server-side media fetchers still enforce their own network guards.
Local media attachments can use absolute paths, workspace-relative paths, or
home-relative ~/ paths. They still pass through the agent file-read policy and
media type checks before delivery.
Valid message-tool payload:
{ "message": "Here is your image.", "mediaUrl": "/workspace/image.png" }
Legacy final assistant reply text may still be normalized for compatibility, but it is not a general plugin/tool protocol.
Plain Markdown image syntax stays text by default. Channels that intentionally
map Markdown image replies to media attachments opt in at their outbound
adapter; Telegram does this so  can still become a media reply.
These directives are separate. Structured media fields and reply/voice tags are
delivery metadata; [embed ...] is the web-only rich render path.
When block streaming is enabled, media must be carried on structured payload fields. If the same media URL is sent in a streamed block and repeated in the final assistant payload, OpenClaw delivers the attachment once and strips the duplicate from the final payload.
[embed ...]
[embed ...] is the only agent-facing rich render syntax for the Control UI.
Self-closing example:
[embed ref="cv_123" title="Status" /]
Rules:
[view ...]is no longer valid for new output.- Embed shortcodes render in the assistant message surface only.
- Only URL-backed embeds are rendered. Use
ref="..."orurl="...". - Block-form inline HTML embed shortcodes are not rendered.
- The web UI strips the shortcode from visible text and renders the embed inline.
- Structured media is not an embed alias and should not be used for rich embed rendering.
Stored rendering shape
The normalized/stored assistant content block is a structured canvas item:
{
"type": "canvas",
"preview": {
"kind": "canvas",
"surface": "assistant_message",
"render": "url",
"viewId": "cv_123",
"url": "/__openclaw__/canvas/documents/cv_123/index.html",
"title": "Status",
"preferredHeight": 320
}
}
Stored/rendered rich blocks use this canvas shape directly. present_view is not recognized.