Web Machine Learning Community Group - Activity
No matches.
Past meetings
-
1 meeting.
Participants
-
9 participants joined.
-
Gaurang Tandon joined.
-
Liad Yosef joined.
-
Aleksei Aleinikov joined.
-
Daniel Mantei joined.
-
song zhang joined.
-
Francesco Marinoni Moretto joined.
-
Artur Serra (Snyk) joined.
-
Vasanth Krishnamoorthy (Pinterest) joined.
-
Christopher Wood (Cloudflare) joined.
-
-
4 participants left.
-
Josi Rosenfeld (CleverThis, Inc.) left.
-
Adnaan Nazir (CleverThis, Inc.) left.
-
Aditya Chhabra (CleverThis, Inc.) left.
-
Vasanth Krishnamoorthy (Pinterest) left.
-
-
4 organizations joined.
-
Level Access joined.
-
Snyk joined.
-
Pinterest joined.
-
Cloudflare joined.
-
-
1 organization left.
-
Pinterest left.
-
GitHub
webmachinelearning/webmcp
-
14 issues opened.
-
Security considerations: tool text can contain characters nobody sees webmachinelearning/webmcp#360, by @kenchris.
-
[Proposal] Soft maximum field length limits for input schema webmachinelearning/webmcp#359, by @jpagnucco.
-
Hosts that load tools lazily: what they need to know, and what pages should not do to avoid the cost webmachinelearning/webmcp#357, by @kenchris.
-
Threat(?): no visibility into the criteria for invoking explicit user confirmation webmachinelearning/webmcp#356, by @BDADAD-stack.
-
Reserve an extension member on tools and results, like MCP's `_meta` webmachinelearning/webmcp#354, by @kenchris.
-
Say how an agent should weigh annotations that the page wrote about itself webmachinelearning/webmcp#353, by @kenchris.
-
Site feedback tools: agents report where a page failed them, and learn known problems first webmachinelearning/webmcp#351, by @siddhi-bansal.
-
Allow `executeTool()` to execute tools registered by opaque-origin documents webmachinelearning/webmcp#349, by @domfarolino.
-
Use the same parallel queue for parallel tasks that represent the same "browser process" webmachinelearning/webmcp#346, by @domfarolino.
-
Don't start a tool whose `executeTool()` call was already cancelled webmachinelearning/webmcp#344, by @beaufortfrancois.
-
Clarify backend-inference, browser-executed tools as a WebMCP use case webmachinelearning/webmcp#343, by @erdalkaraca.
-
Declarative tool-manifest for JS-callable tools: shape proposal, with production reference webmachinelearning/webmcp#341, by @jdkim.
-
Proposal: site guide for agents — a declarative map of screens, tools and workflows (pre-navigation, complements page-scoped tools) webmachinelearning/webmcp#340, by @seunghan91.
-
Fire `toolactivated` after invoking `execute` webmachinelearning/webmcp#332, by @beaufortfrancois.
-
-
3 issues closed.
-
Don't start a tool whose `executeTool()` call was already cancelled webmachinelearning/webmcp#344.
-
Hosting official polyfill for WebMCP webmachinelearning/webmcp#252.
-
Define specific errors for tool execution failures webmachinelearning/webmcp#323.
-
-
34 issues received 76 new comments.
-
Signaling tool stability to agents webmachinelearning/webmcp#328 (11 comments by @dhrumil83, @domfarolino, @kenchris, @khushalsagar, @markafoltz, @sjdallst2)
-
Rename `execute` to `run` webmachinelearning/webmcp#158 (5 comments by @beaufortfrancois, @captainbrosset, @domfarolino, @jakearchibald)
-
Cross-document tool response mechanism webmachinelearning/webmcp#135 (5 comments by @anssiko, @domfarolino, @markafoltz, @sushraja-msft)
-
Scope of `consequentialHint` for high-risk read tools webmachinelearning/webmcp#331 (4 comments by @johannhof, @kenchris, @Kulikowski, @seunghan91)
-
Keeping the agent in-sync with the page's context webmachinelearning/webmcp#312 (4 comments by @beaufortfrancois, @kenchris, @victorhuangwq)
-
Don't start a tool whose `executeTool()` call was already cancelled webmachinelearning/webmcp#344 (4 comments by @beaufortfrancois, @domfarolino)
-
Supporting interoperable probabilistic defense structures through shared attack eval datasets webmachinelearning/webmcp#106 (3 comments by @jpagnucco, @kenchris)
-
No structured way to signal a tool's refusal, distinct from success or a schema-validation error webmachinelearning/webmcp#282 (3 comments by @domfarolino, @kenchris)
-
Introduce semantic hints for side-effects of tools webmachinelearning/webmcp#53 (3 comments by @kenchris)
-
Who owns the validation layer? webmachinelearning/webmcp#92 (3 comments by @beaufortfrancois, @markafoltz, @sjdallst2)
-
Mitigations for profound privacy and security issues seem to be missing? webmachinelearning/webmcp#313 (3 comments by @domfarolino, @jpagnucco, @npdoty)
-
Support multimodal tool input/output webmachinelearning/webmcp#41 (2 comments by @kenchris, @srirammk0)
-
Declarative tool-manifest for JS-callable tools: shape proposal, with production reference webmachinelearning/webmcp#341 (2 comments by @aymen94, @jdkim)
-
Fire `toolactivated` after invoking `execute` webmachinelearning/webmcp#332 (2 comments by @beaufortfrancois)
-
Proposal: Standard MCP-to-WebMCP tool format converter webmachinelearning/webmcp#84 (2 comments by @aymen94, @manticorevault)
-
WebMCP accessibility review checklist webmachinelearning/webmcp#272 (2 comments by @a2uicatalog, @anssiko)
-
[Proposal] Soft maximum field length limits for input schema webmachinelearning/webmcp#359 (1 comment by @kenchris)
-
Hosts that load tools lazily: what they need to know, and what pages should not do to avoid the cost webmachinelearning/webmcp#357 (1 comment by @kenchris)
-
Threat(?): no visibility into the criteria for invoking explicit user confirmation webmachinelearning/webmcp#356 (1 comment by @kenchris)
-
Tool collections: coarse-grained grouping with progressive disclosure webmachinelearning/webmcp#255 (1 comment by @kenchris)
-
WebExtensions API to enumerate and invoke tools webmachinelearning/webmcp#74 (1 comment by @kenchris)
-
WebMCP loses important context when tools appear or disappear webmachinelearning/webmcp#262 (1 comment by @kenchris)
-
WebMCP elicitation via requestUserInteraction() (was: Human in the Loop support for non-browser clients) webmachinelearning/webmcp#165 (1 comment by @kenchris)
-
ToolAnnotations (or something similar) should be implemented for Declarative Tools webmachinelearning/webmcp#163 (1 comment by @Wolfe-Jam)
-
General process setup and collaboration with MCP webmachinelearning/webmcp#28 (1 comment by @Tiancheng-Xu)
-
Define specific errors for tool execution failures webmachinelearning/webmcp#323 (1 comment by @domfarolino)
-
Prevent WebMCP from becoming a browser-dependent substitute for remote MCP webmachinelearning/webmcp#322 (1 comment by @AdamCoulterOz)
-
registerTool should return registered tool instead of undefined webmachinelearning/webmcp#234 (1 comment by @seunghan91)
-
Guidance for sensitive form fields (credentials, card data, national IDs) in declarative and imperative tools webmachinelearning/webmcp#316 (1 comment by @johannhof)
-
-
15 pull requests opened.
-
Say that a hint a tool leaves out is not an assurance webmachinelearning/webmcp#358, by @kenchris.
-
Clarify annotation defaults and add openWorldHint webmachinelearning/webmcp#355, by @kenchris.
-
Add new editor webmachinelearning/webmcp#352, by @anssiko.
-
Editorial: sync explainers with spec (ToolAnnotations link, toolcancel event) webmachinelearning/webmcp#350, by @LF3551.
-
Introduce the `Tool` interface webmachinelearning/webmcp#348, by @domfarolino.
-
Handle same-name re-registration during input serialization webmachinelearning/webmcp#347, by @domfarolino.
-
Fix inactive document unregistration's notification algorithm webmachinelearning/webmcp#345, by @domfarolino.
-
Create makefile.yml webmachinelearning/webmcp#342, by @berhanyosef3-star.
-
Replace UnknownError with specific DOMException types. webmachinelearning/webmcp#339, by @markafoltz.
-
Remove declarative API from spec webmachinelearning/webmcp#338, by @beaufortfrancois.
-
Fix concurrency, detached document, and permissions checks in algorithms. webmachinelearning/webmcp#337, by @markafoltz.
-
Allow executeTool() to return null. webmachinelearning/webmcp#336, by @markafoltz.
-
Fix miscellaneous typos, cross-references, and biblio placement. webmachinelearning/webmcp#335, by @markafoltz.
-
Discuss general agent ecosystem risks and link to agent security guidance webmachinelearning/webmcp#334, by @jpagnucco.
-
Document page-enforced boundaries for write tools webmachinelearning/webmcp#333, by @beladevo.
-
-
10 pull requests merged.
-
Discuss general agent ecosystem risks and link to agent security guidance webmachinelearning/webmcp#334.
-
Add new editor webmachinelearning/webmcp#352.
-
Editorial: sync explainers with spec (ToolAnnotations link, toolcancel event) webmachinelearning/webmcp#350.
-
Handle same-name re-registration during input serialization webmachinelearning/webmcp#347.
-
Fix inactive document unregistration's notification algorithm webmachinelearning/webmcp#345.
-
Fix concurrency, detached document, and permissions checks in algorithms. webmachinelearning/webmcp#337.
-
Fix miscellaneous typos, cross-references, and biblio placement. webmachinelearning/webmcp#335.
-
Replace UnknownError with specific DOMException types. webmachinelearning/webmcp#339.
-
Remove declarative API from spec webmachinelearning/webmcp#338.
-
Add explainer for WebMCP continuations. webmachinelearning/webmcp#327.
-
-
3 pull requests closed.
-
Create makefile.yml webmachinelearning/webmcp#342.
-
Allow executeTool() to return null. webmachinelearning/webmcp#336.
-
Input validation (Fixes #92) webmachinelearning/webmcp#289.
-
-
17 pull requests received 146 new comments.
-
Add `outputSchema` to `ModelContextTool` and `RegisteredTool` webmachinelearning/webmcp#254 (32 comments by @beaufortfrancois, @domfarolino, @kenchris, @markafoltz, @sjdallst2)
-
Add explainer for WebMCP continuations. webmachinelearning/webmcp#327 (23 comments by @beaufortfrancois, @domfarolino, @markafoltz)
-
Introduce the `Tool` interface webmachinelearning/webmcp#348 (21 comments by @domfarolino, @kenchris, @markafoltz)
-
Fix concurrency, detached document, and permissions checks in algorithms. webmachinelearning/webmcp#337 (14 comments by @domfarolino, @markafoltz)
-
Discuss general agent ecosystem risks and link to agent security guidance webmachinelearning/webmcp#334 (10 comments by @domfarolino, @jpagnucco, @markafoltz)
-
Document cross-origin tool exposure and two-sided opt-in in Section 6.3.4 webmachinelearning/webmcp#326 (7 comments by @jpagnucco, @markafoltz, @victorhuangwq)
-
Replace UnknownError with specific DOMException types. webmachinelearning/webmcp#339 (7 comments by @domfarolino, @markafoltz)
-
Fix inactive document unregistration's notification algorithm webmachinelearning/webmcp#345 (5 comments by @beaufortfrancois, @domfarolino)
-
Remove declarative API from spec webmachinelearning/webmcp#338 (5 comments by @beaufortfrancois, @domfarolino)
-
Allow executeTool() to return null. webmachinelearning/webmcp#336 (5 comments by @beaufortfrancois, @domfarolino, @markafoltz)
-
Add Meta Ray-Ban Display to implementation status webmachinelearning/webmcp#325 (4 comments by @anssiko, @beaufortfrancois)
-
Define input-length mitigations in code points webmachinelearning/webmcp#265 (3 comments by @jpagnucco, @kenchris)
-
Document page-enforced boundaries for write tools webmachinelearning/webmcp#333 (3 comments by @domfarolino, @kenchris)
-
Clarify annotation defaults and add openWorldHint webmachinelearning/webmcp#355 (2 comments by @jpagnucco, @kenchris)
-
Handle same-name re-registration during input serialization webmachinelearning/webmcp#347 (2 comments by @beaufortfrancois, @domfarolino)
-
Fix miscellaneous typos, cross-references, and biblio placement. webmachinelearning/webmcp#335 (2 comments by @domfarolino, @markafoltz)
-
webmachinelearning/awesome-webnn
-
1 pull request opened.
-
Add Maxim's Web AI Bench webmachinelearning/awesome-webnn#21, by @ibelem.
-
webmachinelearning/meetings
-
1 issue received 2 new comments.
-
WebML WG/CG F2F Agenda - TPAC 2026 (Dublin, Ireland) webmachinelearning/meetings#39 (2 comments by @anssiko, @KenjiBaheux)
-
-
1 pull request received 1 new comment.
-
Add AGENTS.md for W3C meeting minutes generation webmachinelearning/meetings#40 (1 comment by @ruoxiran)
-
webmachinelearning/webmcp-polyfill
-
1 issue closed.
-
npm package alias webmachinelearning/webmcp-polyfill#5.
-
-
1 issue received 1 new comment.
-
8 pull requests opened.
-
Set copyright attribution to W3C webmachinelearning/webmcp-polyfill#13, by @MiguelsPizza.
-
[WIP] Reject executeTool() failures with the draft's DOMException names webmachinelearning/webmcp-polyfill#12, by @Kulikowski.
-
Relicense under Apache-2.0 webmachinelearning/webmcp-polyfill#11, by @MiguelsPizza.
-
docs(polyfill): drop hand-copied results from TESTING.md webmachinelearning/webmcp-polyfill#10, by @MiguelsPizza.
-
Add declarative tools webmachinelearning/webmcp-polyfill#9, by @MiguelsPizza.
-
Check executeTool() input before reading options webmachinelearning/webmcp-polyfill#8, by @MiguelsPizza.
-
Add ToolActivatedEvent and ToolCancelEvent webmachinelearning/webmcp-polyfill#7, by @beaufortfrancois.
-
Remove origin-keyed agent cluster requirement and update WPT webmachinelearning/webmcp-polyfill#6, by @beaufortfrancois.
-
-
3 pull requests merged.
-
Relicense under Apache-2.0 webmachinelearning/webmcp-polyfill#11.
-
docs(polyfill): drop hand-copied results from TESTING.md webmachinelearning/webmcp-polyfill#10.
-
Remove origin-keyed agent cluster requirement and update WPT webmachinelearning/webmcp-polyfill#6.
-
-
2 pull requests received 11 new comments.
-
[WIP] Reject executeTool() failures with the draft's DOMException names webmachinelearning/webmcp-polyfill#12 (10 comments by @beaufortfrancois, @Kulikowski)
-
webmachinelearning/awesome-webmcp
-
6 pull requests opened.
-
Add Cloud Tools (Websites) and WebMCP Explained (Blogs) webmachinelearning/awesome-webmcp#67, by @LF3551.
-
Add AgentSource to Websites webmachinelearning/awesome-webmcp#66, by @whwebeb65-del.
-
Add Senro to Tools section webmachinelearning/awesome-webmcp#65, by @nartmadi.
-
Add Digital Bricks to Websites webmachinelearning/awesome-webmcp#64, by @raywill.
-
Update webmcp-go and webmcp-django; add webmcp (Ruby) and webmcp-rust to Frameworks webmachinelearning/awesome-webmcp#63, by @seunghan91.
-
Add: Agentic Web Check webmachinelearning/awesome-webmcp#62, by @ericovirgy.
-
-
3 pull requests merged.
-
Add 1 Million Emojis to Websites webmachinelearning/awesome-webmcp#60.
-
Update webmcp-go and webmcp-django; add webmcp (Ruby) and webmcp-rust to Frameworks webmachinelearning/awesome-webmcp#63.
-
Add: Agentic Web Check webmachinelearning/awesome-webmcp#62.
-
-
2 pull requests received 5 new comments.
-
Add 1 Million Emojis to Websites webmachinelearning/awesome-webmcp#60 (4 comments by @anssiko, @cwdx)
-
webmachinelearning/prompt-api
-
2 issues opened.
-
Availability API does not provide actionable failure information webmachinelearning/prompt-api#223, by @sushraja-msft.
-
Thinking effort API - includeThoughts webmachinelearning/prompt-api#222, by @sushraja-msft.
-
-
2 issues closed.
-
Thinking effort API - includeThoughts webmachinelearning/prompt-api#222.
-
Thinking effort API - level options webmachinelearning/prompt-api#220.
-
-
4 issues received 8 new comments.
-
Thinking effort API - includeThoughts webmachinelearning/prompt-api#222 (2 comments by @isaacahouma, @michaelwasserman)
-
Availability API does not provide actionable failure information webmachinelearning/prompt-api#223 (2 comments by @michaelwasserman, @tomayac)
-
Thinking effort API - level options webmachinelearning/prompt-api#220 (2 comments by @michaelwasserman, @sushraja-msft)
-
Consider adding a "fastest" LanguageModelSamplingMode value webmachinelearning/prompt-api#219 (2 comments by @isaacahouma, @reillyeon)
-
-
2 pull requests opened.
-
Give samplingMode precedence over legacy topK and temperature webmachinelearning/prompt-api#225, by @isaacahouma.
-
Update Thinking mode explainer to use thinkingEffort and expectedOutputs webmachinelearning/prompt-api#224, by @isaacahouma.
-
-
3 pull requests merged.
-
Give samplingMode precedence over legacy topK and temperature webmachinelearning/prompt-api#225.
-
Update Thinking mode explainer to use thinkingEffort and expectedOutputs webmachinelearning/prompt-api#224.
-
tool use without auto execution webmachinelearning/prompt-api#162.
-
-
2 pull requests received 40 new comments.
-
tool use without auto execution webmachinelearning/prompt-api#162 (38 comments by @jingyun19, @michaelwasserman, @reillyeon, @sushraja-msft)
-
Update Thinking mode explainer to use thinkingEffort and expectedOutputs webmachinelearning/prompt-api#224 (2 comments by @isaacahouma, @mdrajuhossain848-wq)
-
webmachinelearning/awesome-built-in-ai
-
2 issues opened.
-
[Add Resource] Japan Digital Agency: Natural-Language Public Data Analysis with MCP & Prompt API webmachinelearning/awesome-built-in-ai#11, by @KenjiBaheux.
-
[Add Resource] CSV.it: Client-Side Text-to-CSV Extractor using the Prompt API webmachinelearning/awesome-built-in-ai#10, by @KenjiBaheux.
-
-
1 pull request opened.
-
Add Subtitle Translator webmachinelearning/awesome-built-in-ai#9, by @noffybtw.
-
-
1 pull request received 1 new comment.
webmachinelearning/webnn-docs
-
11 pull requests opened.
-
Add more WebNN demos webmachinelearning/webnn-docs#86, by @ibelem.
-
Update browser-compatibility API and device selection webmachinelearning/webnn-docs#85, by @ibelem.
-
Update architecture chart webmachinelearning/webnn-docs#84, by @ibelem.
-
Update target webmachinelearning/webnn-docs#83, by @ibelem.
-
Package update webmachinelearning/webnn-docs#82, by @ibelem.
-
updates for /learn and /faq webmachinelearning/webnn-docs#81, by @ibelem.
-
api-reference update webmachinelearning/webnn-docs#80, by @ibelem.
-
Revert to 21f804a webmachinelearning/webnn-docs#79, by @ibelem.
-
Fix the build error webmachinelearning/webnn-docs#78, by @ibelem.
-
Fix: Pagefind was not able to build an index webmachinelearning/webnn-docs#77, by @ibelem.
-
package.json update webmachinelearning/webnn-docs#76, by @ibelem.
-
-
11 pull requests merged.
-
Add more WebNN demos webmachinelearning/webnn-docs#86.
-
Update browser-compatibility API and device selection webmachinelearning/webnn-docs#85.
-
Update architecture chart webmachinelearning/webnn-docs#84.
-
Update target webmachinelearning/webnn-docs#83.
-
Package update webmachinelearning/webnn-docs#82.
-
updates for /learn and /faq webmachinelearning/webnn-docs#81.
-
api-reference update webmachinelearning/webnn-docs#80.
-
Revert to 21f804a webmachinelearning/webnn-docs#79.
-
Fix the build error webmachinelearning/webnn-docs#78.
-
Fix: Pagefind was not able to build an index webmachinelearning/webnn-docs#77.
-
package.json update webmachinelearning/webnn-docs#76.
-
webmachinelearning/translation-api
-
1 pull request opened.
-
Fix #summarizer-algorithm reference webmachinelearning/translation-api#88, by @michaelwasserman.
-
-
1 pull request merged.
-
Fix #summarizer-algorithm reference webmachinelearning/translation-api#88.
-
-
1 pull request received 1 new comment.
-
webmachinelearning/translation-api#87 (1 comment by @michaelwasserman)
-
webmachinelearning/webnn-samples-test-framework
-
2 pull requests opened.
-
Monitor uncaught exception in developer preview demos webmachinelearning/webnn-samples-test-framework#70, by @zhibo1-wang.
-
Text generation: reword phi4mini questions webmachinelearning/webnn-samples-test-framework#69, by @zhibo1-wang.
-
-
1 pull request merged.
-
Text generation: reword phi4mini questions webmachinelearning/webnn-samples-test-framework#69.
-
-
1 pull request received 2 new comments.
-
Text generation: reword phi4mini questions webmachinelearning/webnn-samples-test-framework#69 (2 comments by @ibelem, @zhibo1-wang)
-
webmachinelearning/writing-assistance-apis
-
1 issue opened.
-
Add contextUsage and contextWindow to match Prompt API webmachinelearning/writing-assistance-apis#109, by @michaelwasserman.
-
webmachinelearning/proposals
-
1 issue received 1 new comment.
webmachinelearning/webmcp-types
-
1 pull request opened.
-
feat: add agentInvoked and respondWith() to SubmitEvent webmachinelearning/webmcp-types#15, by @MiguelsPizza.
-
webmachinelearning/proofreader-api
-
1 issue received 1 new comment.
-
Dev interest in using the Proofreader API webmachinelearning/proofreader-api#46 (1 comment by @humaknlght)
-
Past meetings
Participants
-
25 participants joined.
-
Nathan Schloss (Meta) joined.
-
Samuel Dallstream (Microsoft Corporation) joined.
-
Melvin Carvalho joined.
-
Idan Levin (nekuda) joined.
-
Eyup Ucmaz joined.
-
Sohail Hooda joined.
-
Ian Holmes joined.
-
Philippe Le Hegaret (W3C) joined.
-
Fucai Xie joined.
-
Pascal Euhus joined.
-
Joe Lamyman (TetraLogical Services Ltd) joined.
-
Renier du Plessis (Agentic Tech LLC) joined.
-
Andrei Nicolae Besleaga joined.
-
Neil Sotirakopoulos joined.
-
Karl Bjorn joined.
-
James (Wolfe) Harrison joined.
-
Evgenii Arsentev (ARSENTEV.AI) joined.
-
Basil Ahamed joined.
-
Nart Madi (Senro) joined.
-
Yunhao Luo (Huawei Canada) joined.
-
Ruoxi Ran (W3C) joined.
-
Ariel Smoliar joined.
-
Nadia Eldeib (CodeYam (Nod Labs, Inc.)) joined.
-
Arnab Mandal Arnab joined.
-
Mahesh Lambe joined.
-
-
4 participants left.
-
Channy Yun left.
-
Grace Snow (TetraLogical Services Ltd) left.
-
Joe Lamyman (TetraLogical Services Ltd) left.
-
Léonie Watson (TetraLogical Services Ltd) left.
-
-
9 organizations joined.
-
nekuda joined.
-
LogMeIn Inc. joined.
-
Reservix GmbH joined.
-
TetraLogical Services Ltd joined.
-
Agentic Tech LLC joined.
-
ARSENTEV.AI joined.
-
Senro joined.
-
Huawei Canada joined.
-
CodeYam (Nod Labs, Inc.) joined.
-
-
1 organization left.
-
TetraLogical Services Ltd left.
-
GitHub
webmachinelearning/prompt-api
-
2 issues opened.
-
Thinking effort API - level options webmachinelearning/prompt-api#220, by @sushraja-msft.
-
Consider adding a "fastest" LanguageModelSamplingMode value webmachinelearning/prompt-api#219, by @isaacahouma.
-
-
5 issues received 7 new comments.
-
Thinking effort API - level options webmachinelearning/prompt-api#220 (3 comments by @isaacahouma, @michaelwasserman, @tomayac)
-
Consider adding a "fastest" LanguageModelSamplingMode value webmachinelearning/prompt-api#219 (1 comment by @tomayac)
-
Tool-calling: would output schemas be useful? webmachinelearning/prompt-api#137 (1 comment by @tomayac)
-
Extend the Message format to support tool calling, extended thinking, and multimodality webmachinelearning/prompt-api#209 (1 comment by @nico-martin)
-
-
2 pull requests opened.
-
Update spec algorithms to include samplingMode webmachinelearning/prompt-api#218, by @isaacahouma.
-
Add Thinking mode section to the Prompt API explainer webmachinelearning/prompt-api#217, by @isaacahouma.
-
-
2 pull requests merged.
-
Update spec algorithms to include samplingMode webmachinelearning/prompt-api#218.
-
Add Thinking mode section to the Prompt API explainer webmachinelearning/prompt-api#217.
-
-
3 pull requests received 52 new comments.
-
tool use without auto execution webmachinelearning/prompt-api#162 (28 comments by @jingyun19, @michaelwasserman, @reillyeon, @tomayac)
-
Add Thinking mode section to the Prompt API explainer webmachinelearning/prompt-api#217 (15 comments by @anssiko, @isaacahouma, @michaelwasserman, @reillyeon, @tomayac)
-
Update spec algorithms to include samplingMode webmachinelearning/prompt-api#218 (9 comments by @isaacahouma, @michaelwasserman, @reillyeon)
-
webmachinelearning/webmcp
-
33 issues opened.
-
Scope of `consequentialHint` for high-risk read tools webmachinelearning/webmcp#331, by @beaufortfrancois.
-
Allow pages to send messages to agents webmachinelearning/webmcp#329, by @Idan-Levin.
-
Signaling tool stability to agents webmachinelearning/webmcp#328, by @domfarolino.
-
Define specific errors for tool execution failures webmachinelearning/webmcp#323, by @domfarolino.
-
Prevent WebMCP from becoming a browser-dependent substitute for remote MCP webmachinelearning/webmcp#322, by @AdamCoulterOz.
-
Deployment data for implementation-status: 580 sites register WebMCP tools at runtime, most of it a handful of plugins webmachinelearning/webmcp#319, by @deangoodman.
-
Origin-Agent-Cluster requirement requires custom server headers on non-Chromium browsers webmachinelearning/webmcp#318, by @beaufortfrancois.
-
Host-controlled WebMCP access between iframes webmachinelearning/webmcp#317, by @sparkinson.
-
Guidance for sensitive form fields (credentials, card data, national IDs) in declarative and imperative tools webmachinelearning/webmcp#316, by @prasanna-gyde.
-
visible UI and forms might have fewer open-ended prompt injection attacks webmachinelearning/webmcp#314, by @npdoty.
-
Mitigations for profound privacy and security issues seem to be missing? webmachinelearning/webmcp#313, by @npdoty.
-
Keeping the agent in-sync with the page's context webmachinelearning/webmcp#312, by @yoavweiss.
-
Preserve what a tool already produced across the execution boundary (unifies #282, #299, #300) webmachinelearning/webmcp#308, by @AshrafAhmed9.
-
Clarify caller lifecycle for non-autosubmit declarative tools webmachinelearning/webmcp#307, by @rgarcia.
-
Proposal: Support Tool Namespaces/Groups for Filtering and Preventing Tool Explosion in Large SPAs webmachinelearning/webmcp#306, by @jackdizhu.
-
Run executeTool completion steps in parallel webmachinelearning/webmcp#305, by @emecii.
-
Implementation information: DeepDeck WebMCP browser and agent webmachinelearning/webmcp#304, by @jo32.
-
Tool aliases webmachinelearning/webmcp#303, by @nartmadi.
-
Clarify that unregistration must not fail an in-flight execution webmachinelearning/webmcp#300, by @minjikim89.
-
Let a tool deliver a final result on caller abort webmachinelearning/webmcp#299, by @minjikim89.
-
§6.4: add page-enforced write boundaries as a mitigation for agent over-reach webmachinelearning/webmcp#298, by @minjikim89.
-
Model-aware tool definitions webmachinelearning/webmcp#297, by @nartmadi.
-
Two agent-browser behaviours a page cannot detect or recover from webmachinelearning/webmcp#293, by @webmyc.
-
Tool results are untyped, so every page invents its own truncation contract webmachinelearning/webmcp#292, by @sreenathmmenon.
-
readOnlyHint cannot distinguish a reversible write from a commit webmachinelearning/webmcp#291, by @sreenathmmenon.
-
Proposal: Focused WebMCP callback observability webmachinelearning/webmcp#290, by @ArielSmoliar.
-
A user agent that both invokes tools and automates the page can complete the page's own human-approval step webmachinelearning/webmcp#288, by @Sidharth1999.
-
Clarification: RegisteredTool.inputSchema / executeTool input / result types — string vs object webmachinelearning/webmcp#287, by @arnabwithab.
-
Include aria-label as part of the Declarative API's toolparamdescription sources webmachinelearning/webmcp#286, by @sawyer-stripe.
-
No structured way to signal a tool's refusal, distinct from success or a schema-validation error webmachinelearning/webmcp#282, by @AshrafAhmed9.
-
Clarify executeTool argument encoding and returned schema shape webmachinelearning/webmcp#278, by @zekariasasaminew.
-
Define accessibility requirements for WebMCP user-agent UI webmachinelearning/webmcp#277, by @mlmrx.
-
Preserve language and direction for declarative tool descriptions webmachinelearning/webmcp#276, by @mlmrx.
-
-
30 issues closed.
-
Origin-Agent-Cluster requirement requires custom server headers on non-Chromium browsers webmachinelearning/webmcp#318.
-
Observations are specified as context thrashing webmachinelearning/webmcp#231.
-
Proposal: Dynamic Tool Definitions in WebMCP webmachinelearning/webmcp#167.
-
Implementation report: agent commerce storefront on WebMCP webmachinelearning/webmcp#266.
-
Implementer note: same-core decision tool + Inspector NL→enum mapping webmachinelearning/webmcp#263.
-
Proposal: preserve completed WebMCP tasks as reviewable workflow documents webmachinelearning/webmcp#261.
-
Proposal: Tool result content types beyond text webmachinelearning/webmcp#86.
-
Run executeTool completion steps in parallel webmachinelearning/webmcp#305.
-
Implementation information: DeepDeck WebMCP browser and agent webmachinelearning/webmcp#304.
-
Proposal: Session and authentication context for tools webmachinelearning/webmcp#87.
-
Skills.sh: A community directory/registry for WebMCP tools (discovery use case) webmachinelearning/webmcp#274.
-
Clarify executeTool argument encoding and returned schema shape webmachinelearning/webmcp#278.
-
Preserve what a tool already produced across the execution boundary (unifies #282, #299, #300) webmachinelearning/webmcp#308.
-
Proposal: [Section 3] Address agent-to-tool trust — identity, scoped permissions, and delegation context webmachinelearning/webmcp#96.
-
Let a tool deliver a final result on caller abort webmachinelearning/webmcp#299.
-
Where to fire `toolactivated` and `toolcanceled` events? webmachinelearning/webmcp#126.
-
visible UI and forms might have fewer open-ended prompt injection attacks webmachinelearning/webmcp#314.
-
Deployment data for implementation-status: 580 sites register WebMCP tools at runtime, most of it a handful of plugins webmachinelearning/webmcp#319.
-
Clarify that unregistration must not fail an in-flight execution webmachinelearning/webmcp#300.
-
Support developer tooling webmachinelearning/webmcp#207.
-
Please formally document Permissions-Policy: tools=() as a security control webmachinelearning/webmcp#178.
-
Tool aliases webmachinelearning/webmcp#303.
-
Proposal: Support Tool Namespaces/Groups for Filtering and Preventing Tool Explosion in Large SPAs webmachinelearning/webmcp#306.
-
Grammar-level (structural) mitigation for prompt injection — field experience from an open spec webmachinelearning/webmcp#239.
-
Two agent-browser behaviours a page cannot detect or recover from webmachinelearning/webmcp#293.
-
Tool results are untyped, so every page invents its own truncation contract webmachinelearning/webmcp#292.
-
readOnlyHint cannot distinguish a reversible write from a commit webmachinelearning/webmcp#291.
-
Proposal: Focused WebMCP callback observability webmachinelearning/webmcp#290.
-
Hint for reversible or consequential actions webmachinelearning/webmcp#176.
-
Clarification: RegisteredTool.inputSchema / executeTool input / result types — string vs object webmachinelearning/webmcp#287.
-
-
57 issues received 150 new comments.
-
Two agent-browser behaviours a page cannot detect or recover from webmachinelearning/webmcp#293 (8 comments by @beaufortfrancois, @domfarolino, @Kulikowski, @webmyc)
-
No structured way to signal a tool's refusal, distinct from success or a schema-validation error webmachinelearning/webmcp#282 (7 comments by @AshrafAhmed9, @domfarolino, @Hronom, @minjikim89, @mlmrx)
-
A user agent that both invokes tools and automates the page can complete the page's own human-approval step webmachinelearning/webmcp#288 (7 comments by @Hronom, @minjikim89, @mlmrx, @Sidharth1999)
-
Tool collections: coarse-grained grouping with progressive disclosure webmachinelearning/webmcp#255 (7 comments by @anssiko, @domfarolino, @jackdizhu, @johnhenry, @jzabroski, @MiguelsPizza)
-
Clarify that unregistration must not fail an in-flight execution webmachinelearning/webmcp#300 (6 comments by @domfarolino, @minjikim89, @mysticalseeker24)
-
WebMCP loses important context when tools appear or disappear webmachinelearning/webmcp#262 (6 comments by @minjikim89, @rohit-binaried, @victorhuangwq, @zekariasasaminew)
-
Origin-Agent-Cluster requirement requires custom server headers on non-Chromium browsers webmachinelearning/webmcp#318 (5 comments by @camillelamy, @domfarolino, @johannhof)
-
Mitigations for profound privacy and security issues seem to be missing? webmachinelearning/webmcp#313 (5 comments by @domfarolino, @Hronom, @jpagnucco, @npdoty)
-
Preserve what a tool already produced across the execution boundary (unifies #282, #299, #300) webmachinelearning/webmcp#308 (5 comments by @domfarolino, @Hronom, @minjikim89, @mlmrx)
-
Let a tool deliver a final result on caller abort webmachinelearning/webmcp#299 (5 comments by @domfarolino, @minjikim89, @mysticalseeker24)
-
§6.4: add page-enforced write boundaries as a mitigation for agent over-reach webmachinelearning/webmcp#298 (5 comments by @Hronom, @minjikim89, @mysticalseeker24, @rohit-binaried)
-
Keeping the agent in-sync with the page's context webmachinelearning/webmcp#312 (4 comments by @beaufortfrancois, @Hronom, @Kulikowski, @rohit-binaried)
-
Frameworks and WebMCP Meta Issue webmachinelearning/webmcp#199 (4 comments by @domfarolino, @rohit-binaried, @sdras)
-
Clarify executeTool argument encoding and returned schema shape webmachinelearning/webmcp#278 (4 comments by @beaufortfrancois, @minjikim89, @zekariasasaminew)
-
Proposal: Dynamic Tool Definitions in WebMCP webmachinelearning/webmcp#167 (4 comments by @arnabwithab, @beaufortfrancois, @domfarolino)
-
Scope of `consequentialHint` for high-risk read tools webmachinelearning/webmcp#331 (3 comments by @domfarolino, @jpagnucco, @victorhuangwq)
-
Proposal: preserve completed WebMCP tasks as reviewable workflow documents webmachinelearning/webmcp#261 (3 comments by @anssiko, @domfarolino, @jo32)
-
Proposal: Session and authentication context for tools webmachinelearning/webmcp#87 (3 comments by @domfarolino, @hemanth)
-
Tool discovery should not be limited to a single traversable navigable webmachinelearning/webmcp#227 (3 comments by @anssiko, @Svishwa2004)
-
WebExtensions API to enumerate and invoke tools webmachinelearning/webmcp#74 (3 comments by @anssiko, @dotproto)
-
Define accessibility requirements for WebMCP user-agent UI webmachinelearning/webmcp#277 (3 comments by @mlmrx, @mysticalseeker24)
-
Proposal: Tool result content types beyond text webmachinelearning/webmcp#86 (2 comments by @domfarolino, @hemanth)
-
Proposal: [Section 3] Address agent-to-tool trust — identity, scoped permissions, and delegation context webmachinelearning/webmcp#96 (2 comments by @domfarolino, @mysticalseeker24)
-
Deployment data for implementation-status: 580 sites register WebMCP tools at runtime, most of it a handful of plugins webmachinelearning/webmcp#319 (2 comments by @deangoodman, @domfarolino)
-
Reposition feature as a generic communication tool? webmachinelearning/webmcp#236 (2 comments by @domfarolino, @johannhof)
-
visible UI and forms might have fewer open-ended prompt injection attacks webmachinelearning/webmcp#314 (2 comments by @domfarolino, @Hronom)
-
WebMCP internationalization review checklist webmachinelearning/webmcp#273 (2 comments by @anssiko, @mlmrx)
-
WebMCP accessibility review checklist webmachinelearning/webmcp#272 (2 comments by @anssiko, @mlmrx)
-
Hosting official polyfill for WebMCP webmachinelearning/webmcp#252 (2 comments by @anssiko, @beaufortfrancois)
-
Proposal: Skills — workflow-level context for tool composition webmachinelearning/webmcp#161 (2 comments by @hungson175, @volkandkaya)
-
readOnlyHint cannot distinguish a reversible write from a commit webmachinelearning/webmcp#291 (2 comments by @domfarolino, @mlmrx)
-
Clarification: RegisteredTool.inputSchema / executeTool input / result types — string vs object webmachinelearning/webmcp#287 (2 comments by @arnabwithab, @mlmrx)
-
Grammar-level (structural) mitigation for prompt injection — field experience from an open spec webmachinelearning/webmcp#239 (2 comments by @johannhof, @victorhuangwq)
-
Include aria-label as part of the Declarative API's toolparamdescription sources webmachinelearning/webmcp#286 (2 comments by @mlmrx, @sawyer-stripe)
-
Consider restricting maximum input lengths webmachinelearning/webmcp#73 (1 comment by @JasonColapietro)
-
Implementation report: agent commerce storefront on WebMCP webmachinelearning/webmcp#266 (1 comment by @domfarolino)
-
Implementer note: same-core decision tool + Inspector NL→enum mapping webmachinelearning/webmcp#263 (1 comment by @domfarolino)
-
Prevent WebMCP from becoming a browser-dependent substitute for remote MCP webmachinelearning/webmcp#322 (1 comment by @domfarolino)
-
Define specific errors for tool execution failures webmachinelearning/webmcp#323 (1 comment by @domfarolino)
-
Implementation information: DeepDeck WebMCP browser and agent webmachinelearning/webmcp#304 (1 comment by @domfarolino)
-
Skills.sh: A community directory/registry for WebMCP tools (discovery use case) webmachinelearning/webmcp#274 (1 comment by @domfarolino)
-
Host-controlled WebMCP access between iframes webmachinelearning/webmcp#317 (1 comment by @domfarolino)
-
Where to fire `toolactivated` and `toolcanceled` events? webmachinelearning/webmcp#126 (1 comment by @beaufortfrancois)
-
Define the declarative input-schema synthesis algorithm (#declarative-api is currently a TODO) webmachinelearning/webmcp#210 (1 comment by @theadamdanielsson)
-
Supporting interoperable probabilistic defense structures through shared attack eval datasets webmachinelearning/webmcp#106 (1 comment by @realArcherL)
-
How to expose failed declarative tool registration webmachinelearning/webmcp#182 (1 comment by @anssiko)
-
Tool results are untyped, so every page invents its own truncation contract webmachinelearning/webmcp#292 (1 comment by @domfarolino)
-
Proposal: Focused WebMCP callback observability webmachinelearning/webmcp#290 (1 comment by @domfarolino)
-
-
24 pull requests opened.
-
Remove origin-keyed agent cluster requirement webmachinelearning/webmcp#330, by @beaufortfrancois.
-
Add explainer for WebMCP continuations. webmachinelearning/webmcp#327, by @markafoltz.
-
Document cross-origin tool exposure and two-sided opt-in in Section 6.3.4 webmachinelearning/webmcp#326, by @jpagnucco.
-
Add Meta Ray-Ban Display to implementation status webmachinelearning/webmcp#325, by @beaufortfrancois.
-
Treat an omitted `executeTool()` input as an empty object webmachinelearning/webmcp#324, by @MiguelsPizza.
-
Fix typos and Markdown formatting in the explainer docs webmachinelearning/webmcp#321, by @lazizbekravshanov.
-
Point the API TODO at the README instead of the removed proposal.md webmachinelearning/webmcp#320, by @lazizbekravshanov.
-
Non-normatively reference the MCP specification webmachinelearning/webmcp#315, by @mconroy-cf.
-
Clarify execution behavior after tool unregistration webmachinelearning/webmcp#311, by @1aifanatic.
-
Document dynamic tool definition patterns webmachinelearning/webmcp#310, by @1aifanatic.
-
Use the implementers spelling, matching the rest of the repo webmachinelearning/webmcp#309, by @lazizbekravshanov.
-
Fix malformed tool registration and lookup steps webmachinelearning/webmcp#302, by @emecii.
-
Fix execution context and navigable lookup algorithms webmachinelearning/webmcp#301, by @emecii.
-
Update README with refined definition of headless usecases webmachinelearning/webmcp#296, by @sdras.
-
Refactor sequence diagrams webmachinelearning/webmcp#295, by @domfarolino.
-
Add Best Practices section and guidance on dynamic tool registration to README webmachinelearning/webmcp#294, by @beaufortfrancois.
-
Input validation (Fixes #92) webmachinelearning/webmcp#289, by @arnabwithab.
-
Fix the JSON-LD media type in the declarative explainer webmachinelearning/webmcp#285, by @lazizbekravshanov.
-
Update stale examples and references to the current modelContext API webmachinelearning/webmcp#284, by @lazizbekravshanov.
-
Update implementation status to include agent support webmachinelearning/webmcp#283, by @beaufortfrancois.
-
Include origin in observation tool map via observed tool collection struct webmachinelearning/webmcp#281, by @johannhof.
-
Add flag for local development in Chrome webmachinelearning/webmcp#280, by @beaufortfrancois.
-
Clarify executeTool input and schema values webmachinelearning/webmcp#279, by @mlmrx.
-
Document Permissions Policy as a security mitigation webmachinelearning/webmcp#275, by @mlmrx.
-
-
26 pull requests merged.
-
Remove origin-keyed agent cluster requirement webmachinelearning/webmcp#330.
-
Clarify observation context management webmachinelearning/webmcp#264.
-
Document dynamic tool definition patterns webmachinelearning/webmcp#310.
-
Treat an omitted `executeTool()` input as an empty object webmachinelearning/webmcp#324.
-
Add Meta Ray-Ban Display to implementation status webmachinelearning/webmcp#325.
-
Point the API TODO at the README instead of the removed proposal.md webmachinelearning/webmcp#320.
-
Non-normatively reference the MCP specification webmachinelearning/webmcp#315.
-
Fix execution context and navigable lookup algorithms webmachinelearning/webmcp#301.
-
Fix typos and Markdown formatting in the explainer docs webmachinelearning/webmcp#321.
-
Clarify executeTool input and schema values webmachinelearning/webmcp#279.
-
Clarify execution behavior after tool unregistration webmachinelearning/webmcp#311.
-
Add ToolActivatedEvent and ToolCancelEvent interfaces webmachinelearning/webmcp#245.
-
Add `debugging` member to ToolAnnotations webmachinelearning/webmcp#253.
-
Document Permissions Policy as a security mitigation webmachinelearning/webmcp#275.
-
Fix malformed tool registration and lookup steps webmachinelearning/webmcp#302.
-
Use the implementers spelling, matching the rest of the repo webmachinelearning/webmcp#309.
-
Change `executeTool()` inputObject type from `object` to `any` webmachinelearning/webmcp#251.
-
Include origin in observation tool map via observed tool collection struct webmachinelearning/webmcp#281.
-
Update README with refined definition of headless usecases webmachinelearning/webmcp#296.
-
Add flag for local development in Chrome webmachinelearning/webmcp#280.
-
Update stale examples and references to the current modelContext API webmachinelearning/webmcp#284.
-
Fix the JSON-LD media type in the declarative explainer webmachinelearning/webmcp#285.
-
Add Best Practices section and guidance on dynamic tool registration to README webmachinelearning/webmcp#294.
-
Refactor sequence diagrams webmachinelearning/webmcp#295.
-
Add consequentialHint to ToolAnnotations webmachinelearning/webmcp#217.
-
Update implementation status to include agent support webmachinelearning/webmcp#283.
-
-
26 pull requests received 94 new comments.
-
Add explainer for WebMCP continuations. webmachinelearning/webmcp#327 (16 comments by @beaufortfrancois, @domfarolino, @markafoltz)
-
Treat an omitted `executeTool()` input as an empty object webmachinelearning/webmcp#324 (9 comments by @anssiko, @beaufortfrancois, @domfarolino, @MiguelsPizza)
-
Change `executeTool()` inputObject type from `object` to `any` webmachinelearning/webmcp#251 (9 comments by @beaufortfrancois, @domfarolino)
-
Fix malformed tool registration and lookup steps webmachinelearning/webmcp#302 (8 comments by @beaufortfrancois, @emecii)
-
Clarify executeTool input and schema values webmachinelearning/webmcp#279 (6 comments by @domfarolino)
-
Document Permissions Policy as a security mitigation webmachinelearning/webmcp#275 (6 comments by @domfarolino, @mlmrx)
-
Drop the stale native-agents note from the security questionnaire webmachinelearning/webmcp#269 (4 comments by @lazizbekravshanov, @victorhuangwq)
-
Non-normatively reference the MCP specification webmachinelearning/webmcp#315 (4 comments by @anssiko, @mconroy-cf, @w3cbot)
-
Include origin in observation tool map via observed tool collection struct webmachinelearning/webmcp#281 (3 comments by @domfarolino, @johannhof)
-
Add flag for local development in Chrome webmachinelearning/webmcp#280 (3 comments by @beaufortfrancois, @domfarolino)
-
Document cross-origin tool exposure and two-sided opt-in in Section 6.3.4 webmachinelearning/webmcp#326 (2 comments by @anssiko, @victorhuangwq)
-
Define input-length mitigations in code points webmachinelearning/webmcp#265 (2 comments by @domfarolino, @jpagnucco)
-
Fix execution context and navigable lookup algorithms webmachinelearning/webmcp#301 (2 comments by @domfarolino, @emecii)
-
Add `debugging` member to ToolAnnotations webmachinelearning/webmcp#253 (2 comments by @beaufortfrancois, @markafoltz)
-
Add `outputSchema` to `ModelContextTool` and `RegisteredTool` webmachinelearning/webmcp#254 (2 comments by @beaufortfrancois, @volkandkaya)
-
Add Meta Ray-Ban Display to implementation status webmachinelearning/webmcp#325 (1 comment by @beaufortfrancois)
-
Point the API TODO at the README instead of the removed proposal.md webmachinelearning/webmcp#320 (1 comment by @w3cbot)
-
Fix typos and Markdown formatting in the explainer docs webmachinelearning/webmcp#321 (1 comment by @w3cbot)
-
Add ToolActivatedEvent and ToolCancelEvent interfaces webmachinelearning/webmcp#245 (1 comment by @markafoltz)
-
Use the implementers spelling, matching the rest of the repo webmachinelearning/webmcp#309 (1 comment by @w3cbot)
-
Change `RegisteredTool#inputSchema` type from `DOMString` to `object` webmachinelearning/webmcp#241 (1 comment by @Buffalo456)
-
Fix the JSON-LD media type in the declarative explainer webmachinelearning/webmcp#285 (1 comment by @w3cbot)
-
Update stale examples and references to the current modelContext API webmachinelearning/webmcp#284 (1 comment by @lazizbekravshanov)
-
Add consequentialHint to ToolAnnotations webmachinelearning/webmcp#217 (1 comment by @beaufortfrancois)
-
webmachinelearning/awesome-webmcp
-
2 issues opened.
-
Add Webmcp Skills Collection webmachinelearning/awesome-webmcp#58, by @JustineDevs.
-
Suggest Busleyden Free Line Sheet for Websites webmachinelearning/awesome-webmcp#38, by @AlexanderLuzDH.
-
-
2 issues closed.
-
Add Webmcp Skills Collection webmachinelearning/awesome-webmcp#58.
-
Community Tooling for WebMCP: MCP Server Bridge, CDP Library, React Bindings & DevTools Extension webmachinelearning/awesome-webmcp#1.
-
-
3 issues received 4 new comments.
-
Community Tooling for WebMCP: MCP Server Bridge, CDP Library, React Bindings & DevTools Extension webmachinelearning/awesome-webmcp#1 (1 comment by @anssiko)
-
25 pull requests opened.
-
Add Patrick Brosset dotAI WebMCP talk and slides webmachinelearning/awesome-webmcp#61, by @victorhuangwq.
-
Add 1 Million Emojis to Websites webmachinelearning/awesome-webmcp#60, by @cwdx.
-
Add WebMCP Agents skill collection webmachinelearning/awesome-webmcp#59, by @JustineDevs.
-
Add Known Good WebMCP checker to Tools webmachinelearning/awesome-webmcp#57, by @deangoodman.
-
Add Microsoft Edge WebMCP Labs samples and Explorer webmachinelearning/awesome-webmcp#56, by @victorhuangwq.
-
Add Intlayer project to README webmachinelearning/awesome-webmcp#54, by @aymericzip.
-
Add Senro to Tools webmachinelearning/awesome-webmcp#53, by @nartmadi.
-
Add Site Passport to Websites webmachinelearning/awesome-webmcp#52, by @renduples.
-
Add Syntax episode "WebMCP is here (and you should care)" to Videos webmachinelearning/awesome-webmcp#51, by @oddrationale.
-
Add AGENTS.md WebMCP playground to Demos webmachinelearning/awesome-webmcp#50, by @Wolfe-Jam.
-
Add Timeline Studio to Websites webmachinelearning/awesome-webmcp#49, by @MartinDelophy.
-
Add Deveserisso to Websites webmachinelearning/awesome-webmcp#48, by @lpcruz2.
-
Add Site Passport to Tools webmachinelearning/awesome-webmcp#47, by @renduples.
-
Add turva.dev webmachinelearning/awesome-webmcp#46, by @erekola.
-
Add Haul Handbook trucking reference and browser tools webmachinelearning/awesome-webmcp#45, by @ereid7.
-
Add Settled Estate browser WebMCP tools webmachinelearning/awesome-webmcp#44, by @ereid7.
-
Add Parity to Websites webmachinelearning/awesome-webmcp#43, by @mysticalseeker24.
-
Add Signett to Frameworks webmachinelearning/awesome-webmcp#42, by @kartik-hegde.
-
Add scvd.store to Websites webmachinelearning/awesome-webmcp#41, by @cv-scvd.
-
Add GunsNation webmachinelearning/awesome-webmcp#40, by @sneg55.
-
Add Simple Tools to Tools and Websites webmachinelearning/awesome-webmcp#39, by @salvarecuero.
-
Add TrickyBird to Websites webmachinelearning/awesome-webmcp#37, by @trickybird-net.
-
Add CliDeck Network Evidence Workbench demo webmachinelearning/awesome-webmcp#36, by @SmartRoot7.
-
Update Scholar Sidekick entry: seven tools, direct WebMCP page webmachinelearning/awesome-webmcp#35, by @mlava.
-
Add Archipelago to Websites webmachinelearning/awesome-webmcp#32, by @warrenperez.
-
-
29 pull requests merged.
-
Add Patrick Brosset dotAI WebMCP talk and slides webmachinelearning/awesome-webmcp#61.
-
Add WebMCP Agents skill collection webmachinelearning/awesome-webmcp#59.
-
Add Known Good WebMCP checker to Tools webmachinelearning/awesome-webmcp#57.
-
Add Intlayer project to README webmachinelearning/awesome-webmcp#54.
-
Add Site Passport to Websites webmachinelearning/awesome-webmcp#52.
-
Add Syntax episode "WebMCP is here (and you should care)" to Videos webmachinelearning/awesome-webmcp#51.
-
Add AGENTS.md WebMCP playground to Demos webmachinelearning/awesome-webmcp#50.
-
Add Timeline Studio to Websites webmachinelearning/awesome-webmcp#49.
-
Add Settled Estate browser WebMCP tools webmachinelearning/awesome-webmcp#44.
-
Add Parity to Websites webmachinelearning/awesome-webmcp#43.
-
Add Signett to Frameworks webmachinelearning/awesome-webmcp#42.
-
Add Simple Tools to Tools and Websites webmachinelearning/awesome-webmcp#39.
-
Add Orbilo to Tools webmachinelearning/awesome-webmcp#31.
-
Add Microsoft Edge WebMCP Labs samples and Explorer webmachinelearning/awesome-webmcp#56.
-
Add NaCzarter to Websites webmachinelearning/awesome-webmcp#14.
-
Add CliDeck Network Evidence Workbench demo webmachinelearning/awesome-webmcp#36.
-
Update Scholar Sidekick entry: seven tools, direct WebMCP page webmachinelearning/awesome-webmcp#35.
-
Add Archipelago to Websites webmachinelearning/awesome-webmcp#32.
-
Add AIC assurance tooling webmachinelearning/awesome-webmcp#30.
-
Add simple-webmcp webmachinelearning/awesome-webmcp#28.
-
Add WindTunnel and WebMCP Kit webmachinelearning/awesome-webmcp#24.
-
Add WebMCP Today to Tools webmachinelearning/awesome-webmcp#23.
-
Add webmcp-go to Frameworks webmachinelearning/awesome-webmcp#20.
-
Add webmcp-django to Frameworks webmachinelearning/awesome-webmcp#19.
-
Add admintoolkit.io to Tools webmachinelearning/awesome-webmcp#18.
-
Add Scholar Sidekick to Websites webmachinelearning/awesome-webmcp#12.
-
Add WebConverter under Tools webmachinelearning/awesome-webmcp#8.
-
Add agent-ready.dev to Websites webmachinelearning/awesome-webmcp#11.
-
Add webmcp.cool and Ask nekuda to Tools webmachinelearning/awesome-webmcp#9.
-
-
5 pull requests closed.
-
Add Senro to Tools webmachinelearning/awesome-webmcp#53.
-
Add GunsNation webmachinelearning/awesome-webmcp#40.
-
Add turva.dev webmachinelearning/awesome-webmcp#46.
-
Add Site Passport to Tools webmachinelearning/awesome-webmcp#47.
-
Add Haul Handbook trucking reference and browser tools webmachinelearning/awesome-webmcp#45.
-
-
19 pull requests received 24 new comments.
-
Add NaCzarter to Websites webmachinelearning/awesome-webmcp#14 (4 comments by @anssiko, @marczynski)
-
Add TrickyBird to Websites webmachinelearning/awesome-webmcp#37 (2 comments by @anssiko, @trickybird-net)
-
Add Patrick Brosset dotAI WebMCP talk and slides webmachinelearning/awesome-webmcp#61 (1 comment by @victorhuangwq)
-
Add WebMCP Attack Surface Scanner to README webmachinelearning/awesome-webmcp#22 (1 comment by @anssiko)
-
Update Scholar Sidekick entry: seven tools, direct WebMCP page webmachinelearning/awesome-webmcp#35 (1 comment by @mlava)
-
feat: add webmcp-sdk (Tools) and AgentPay x402 payment demo (Demos) webmachinelearning/awesome-webmcp#5 (1 comment by @anssiko)
-
webmachinelearning/translation-api
-
3 issues received 3 new comments.
-
Clarification on Definition of "On-Device Models" Based on TAG Feedback webmachinelearning/translation-api#72 (1 comment by @toreini)
-
What **interoperability guarantees** apply (which markup/attributes are preserved, translated, or may change)? webmachinelearning/translation-api#83 (1 comment by @adarshmishra07)
-
-
1 pull request opened.
-
I added a security text for issue 78 webmachinelearning/translation-api#86, by @toreini.
-
webmachinelearning/proposals
-
2 issues opened.
-
Decision Models webmachinelearning/proposals#20, by @AdamSobieski.
-
Generating Multimedia and AI Agents from Web Content webmachinelearning/proposals#19, by @AdamSobieski.
-
webmachinelearning/awesome-built-in-ai
-
1 issue opened.
-
[Add Resource] Shaka Player: Native Subtitle Translation with On-Device Speech-to-Text combo webmachinelearning/awesome-built-in-ai#8, by @KenjiBaheux.
-
-
4 issues closed.
-
[Add Resource] Semantic Sticky Board: Client-Side Semantic Clustering with Semantic Embedder API webmachinelearning/awesome-built-in-ai#5.
-
Consider adding AAC Board webmachinelearning/awesome-built-in-ai#3.
-
[Add Resource] Rakuten: Client-Side Observability & OpenTelemetry for Prompt API webmachinelearning/awesome-built-in-ai#4.
-
Consider adding our own I/O content and Web AI Summit content webmachinelearning/awesome-built-in-ai#1.
-
-
2 pull requests opened.
-
docs: add new project examples, developer tools, and conference talk playlists to README webmachinelearning/awesome-built-in-ai#7, by @KenjiBaheux.
-
Add Web AI Agent Skills webmachinelearning/awesome-built-in-ai#6, by @webmaxru.
-
-
2 pull requests merged.
-
docs: add new project examples, developer tools, and conference talk playlists to README webmachinelearning/awesome-built-in-ai#7.
-
Add Web AI Agent Skills webmachinelearning/awesome-built-in-ai#6.
-
webmachinelearning/webmcp-polyfill
-
1 issue opened.
-
npm package alias webmachinelearning/webmcp-polyfill#5, by @MiguelsPizza.
-
-
1 issue received 4 new comments.
-
npm package alias webmachinelearning/webmcp-polyfill#5 (4 comments by @43081j, @anssiko, @beaufortfrancois, @MiguelsPizza)
-
-
4 pull requests opened.
-
fix: emit frame bridge when preparing nested Git dependencies webmachinelearning/webmcp-polyfill#4, by @MiguelsPizza.
-
Document and test extension integration webmachinelearning/webmcp-polyfill#3, by @MiguelsPizza.
-
Add executeTool() webmachinelearning/webmcp-polyfill#2, by @MiguelsPizza.
-
Add WebMCP registration, discovery, and execution webmachinelearning/webmcp-polyfill#1, by @MiguelsPizza.
-
-
2 pull requests merged.
-
Add WebMCP registration, discovery, and execution webmachinelearning/webmcp-polyfill#1.
-
Add executeTool() webmachinelearning/webmcp-polyfill#2.
-
-
2 pull requests closed.
-
Document and test extension integration webmachinelearning/webmcp-polyfill#3.
-
fix: emit frame bridge when preparing nested Git dependencies webmachinelearning/webmcp-polyfill#4.
-
-
1 pull request received 31 new comments.
-
Add WebMCP registration, discovery, and execution webmachinelearning/webmcp-polyfill#1 (31 comments by @anssiko, @beaufortfrancois, @MiguelsPizza)
-
webmachinelearning/webnn-samples-test-framework
-
1 pull request opened.
-
sdxl: add test webmachinelearning/webnn-samples-test-framework#68, by @zhibo1-wang.
-
-
1 pull request merged.
-
sdxl: add test webmachinelearning/webnn-samples-test-framework#68.
-
-
1 pull request received 1 new comment.
webmachinelearning/webmcp-types
-
1 issue closed.
-
Automatically infer args from input schema? webmachinelearning/webmcp-types#5.
-
-
4 pull requests opened.
-
feat: add toolactivated and toolcancel events webmachinelearning/webmcp-types#13, by @beaufortfrancois.
-
feat: add debugging to ToolAnnotations webmachinelearning/webmcp-types#11, by @beaufortfrancois.
-
feat: add consequentialHint to ToolAnnotations webmachinelearning/webmcp-types#8, by @beaufortfrancois.
-
chore: automate npm releases with Release Please and trusted publishing webmachinelearning/webmcp-types#7, by @MiguelsPizza.
-
-
6 pull requests merged.
-
feat: add toolactivated and toolcancel events webmachinelearning/webmcp-types#13.
-
feat: add debugging to ToolAnnotations webmachinelearning/webmcp-types#11.
-
Add executeTool() method webmachinelearning/webmcp-types#3.
-
feat: add consequentialHint to ToolAnnotations webmachinelearning/webmcp-types#8.
-
chore: automate npm releases with Release Please and trusted publishing webmachinelearning/webmcp-types#7.
-
Infer execute input from inputSchema webmachinelearning/webmcp-types#6.
-
-
4 pull requests received 9 new comments.
-
Infer execute input from inputSchema webmachinelearning/webmcp-types#6 (5 comments by @beaufortfrancois, @MiguelsPizza)
-
Add executeTool() method webmachinelearning/webmcp-types#3 (2 comments by @beaufortfrancois, @MiguelsPizza)
-
webmachinelearning/webmcp-types#9 (1 comment by @beaufortfrancois)
-
chore: automate npm releases with Release Please and trusted publishing webmachinelearning/webmcp-types#7 (1 comment by @beaufortfrancois)
-
webmachinelearning/webnn-native
-
1 issue closed.
-
[DML] Optimizations to improve DML memory utilization webmachinelearning/webnn-native#221.
-
webmachinelearning/webmachinelearning.github.io
-
1 pull request opened.
-
Add WebMCP section and intro webmachinelearning/webmachinelearning.github.io#114, by @anssiko.
-
-
1 pull request merged.
-
Add WebMCP section and intro webmachinelearning/webmachinelearning.github.io#114.
-
webmachinelearning/meetings
-
1 issue opened.
-
Video annotation for machine learning with WebVMT at TPAC? webmachinelearning/meetings#41, by @rjksmith.
-
-
1 issue received 1 new comment.
-
Video annotation for machine learning with WebVMT at TPAC? webmachinelearning/meetings#41 (1 comment by @anssiko)
-
-
1 pull request opened.
-
Add AGENTS.md for W3C meeting minutes generation webmachinelearning/meetings#40, by @anssiko.
-
webmachinelearning/awesome-webnn
-
1 pull request opened.
-
Sept 2026 update webmachinelearning/awesome-webnn#20, by @ibelem.
-
-
1 pull request merged.
-
Sept 2026 update webmachinelearning/awesome-webnn#20.
-
Past meetings
-
1 meeting.
Participants
-
13 participants joined.
-
Channy Yun joined.
-
Wei Herr Teng joined.
-
Chamara Somaratne (Sovereign Stack Global) joined.
-
Naveen Chatlapalli (Ashling) joined.
-
James Yang joined.
-
Hitesh Parmar joined.
-
Prashastha Mudannayake joined.
-
Josh Inch (Meta) joined.
-
Jay Mok (PayPal) joined.
-
Yifan Chen joined.
-
Yifan Chen joined.
-
Delta Krista joined.
-
Kannadasan Saminathan joined.
-
-
3 participants left.
-
Richard MacManus left.
-
Fabiano Magalhaes left.
-
Yifan Chen left.
-
-
3 organizations joined.
-
Sovereign Stack Global joined.
-
Ashling joined.
-
PayPal joined.
-
GitHub
webmachinelearning/webmcp
-
21 issues opened.
-
Skills.sh: A community directory/registry for WebMCP tools (discovery use case) webmachinelearning/webmcp#274, by @JustineDevs.
-
WebMCP internationalization review checklist webmachinelearning/webmcp#273, by @anssiko.
-
WebMCP accessibility review checklist webmachinelearning/webmcp#272, by @anssiko.
-
Add an optional `group` field to the tool definition for dev‑tools grouping webmachinelearning/webmcp#271, by @baishiwen9.
-
Add an optional `group` field to the tool definition for dev‑tools grouping webmachinelearning/webmcp#270, by @baishiwen9.
-
[bug] Working on Brave, not working on MS Edge and Chrome webmachinelearning/webmcp#268, by @thiagodp.
-
Turn awareness webmachinelearning/webmcp#267, by @yoavweiss.
-
Implementation report: agent commerce storefront on WebMCP webmachinelearning/webmcp#266, by @JasonColapietro.
-
Implementer note: same-core decision tool + Inspector NL→enum mapping webmachinelearning/webmcp#263, by @davewo83.
-
WebMCP loses important context when tools appear or disappear webmachinelearning/webmcp#262, by @heyitischris-web.
-
Proposal: preserve completed WebMCP tasks as reviewable workflow documents webmachinelearning/webmcp#261, by @tabilet.
-
Agent-Scoped Cookies webmachinelearning/webmcp#257, by @Idan-Levin.
-
Origin Trial research note: observations on tool exposure, validation, and context cost webmachinelearning/webmcp#256, by @SeCuReDmE-main-dev.
-
Tool collections: coarse-grained grouping with progressive disclosure webmachinelearning/webmcp#255, by @sdras.
-
Hosting official polyfill for WebMCP webmachinelearning/webmcp#252, by @beaufortfrancois.
-
The `executeTool()` method should take an object, not a string webmachinelearning/webmcp#243, by @domfarolino.
-
App -> Agent explicit integration/manifest? webmachinelearning/webmcp#242, by @pmeenan.
-
Grammar-level (structural) mitigation for prompt injection — field experience from an open spec webmachinelearning/webmcp#239, by @ais-space.
-
ioiopaut789 webmachinelearning/webmcp#238, by @panyachai2018-crypto.
-
Various questions & feedback webmachinelearning/webmcp#237, by @jakearchibald.
-
Reposition feature as a generic communication tool? webmachinelearning/webmcp#236, by @jakearchibald.
-
-
17 issues closed.
-
Add an optional `group` field to the tool definition for dev‑tools grouping webmachinelearning/webmcp#271.
-
Add an optional `group` field to the tool definition for dev‑tools grouping webmachinelearning/webmcp#270.
-
[bug] Working on Brave, not working on MS Edge and Chrome webmachinelearning/webmcp#268.
-
Define the API for in-page Agents to use a site's declared tools webmachinelearning/webmcp#51.
-
Should unregistering a tool cancel an in-flight executions? webmachinelearning/webmcp#218.
-
Add a callback to inform the site when tool invocation should be aborted webmachinelearning/webmcp#48.
-
The `executeTool()` method should take an object, not a string webmachinelearning/webmcp#243.
-
registerTool() stores the tool definition in the model context struct itself, not its tool map webmachinelearning/webmcp#213.
-
perform an observation asserts the active document is not fully active (inverted condition) webmachinelearning/webmcp#214.
-
Redundancy with the accessibility tree webmachinelearning/webmcp#91.
-
Protecting tools from 3P scripts in the top-level context webmachinelearning/webmcp#159.
-
Support cross-frame tool enumeration and composability while handling name collisions webmachinelearning/webmcp#160.
-
Allow iframes to declare tools webmachinelearning/webmcp#57.
-
Allow iframes to deal with tools in their parent/top-level context webmachinelearning/webmcp#117.
-
Hosting official TypeScript types for WebMCP webmachinelearning/webmcp#190.
-
ioiopaut789 webmachinelearning/webmcp#238.
-
-
1 issue reopened.
-
Should unregistering a tool cancel an in-flight executions? webmachinelearning/webmcp#218, by @domfarolino.
-
-
25 issues received 75 new comments.
-
Origin Trial research note: observations on tool exposure, validation, and context cost webmachinelearning/webmcp#256 (9 comments by @heyitischris-web, @NoClueMike, @SeCuReDmE-main-dev)
-
Reposition feature as a generic communication tool? webmachinelearning/webmcp#236 (8 comments by @domfarolino, @jakearchibald, @martinthomson)
-
Transitioning WebMCP to the working group? webmachinelearning/webmcp#192 (7 comments by @anssiko, @domfarolino, @marcoscaceres, @mwyrzykowski)
-
Redundancy with the accessibility tree webmachinelearning/webmcp#91 (6 comments by @cookiecrook, @domfarolino, @Guandaline, @LJWatson)
-
Tool collections: coarse-grained grouping with progressive disclosure webmachinelearning/webmcp#255 (5 comments by @domfarolino, @Idan-Levin, @MiguelsPizza)
-
App -> Agent explicit integration/manifest? webmachinelearning/webmcp#242 (4 comments by @heyitischris-web, @pmeenan)
-
Hosting official polyfill for WebMCP webmachinelearning/webmcp#252 (4 comments by @anssiko, @domfarolino, @igrigorik)
-
Should unregistering a tool cancel an in-flight executions? webmachinelearning/webmcp#218 (4 comments by @beaufortfrancois, @domfarolino)
-
Various questions & feedback webmachinelearning/webmcp#237 (4 comments by @beaufortfrancois, @domfarolino)
-
Grammar-level (structural) mitigation for prompt injection — field experience from an open spec webmachinelearning/webmcp#239 (3 comments by @ais-space, @johannhof)
-
registerTool should return registered tool instead of undefined webmachinelearning/webmcp#234 (3 comments by @beaufortfrancois, @domfarolino, @seunghan91)
-
Proposal: Dynamic Tool Definitions in WebMCP webmachinelearning/webmcp#167 (2 comments by @arnabwithab, @domfarolino)
-
[bug] Working on Brave, not working on MS Edge and Chrome webmachinelearning/webmcp#268 (2 comments by @thiagodp, @victorhuangwq)
-
Tool discovery should not be limited to a single traversable navigable webmachinelearning/webmcp#227 (2 comments by @beaufortfrancois, @domfarolino)
-
Add an optional `group` field to the tool definition for dev‑tools grouping webmachinelearning/webmcp#271 (1 comment by @domfarolino)
-
WebMCP elicitation via requestUserInteraction() (was: Human in the Loop support for non-browser clients) webmachinelearning/webmcp#165 (1 comment by @thankywal)
-
Implementation report: agent commerce storefront on WebMCP webmachinelearning/webmcp#266 (1 comment by @JasonColapietro)
-
Proposal: API for connecting to remote MCP servers webmachinelearning/webmcp#83 (1 comment by @fredericbarthelet)
-
Future length limits need to define the text unit webmachinelearning/webmcp#219 (1 comment by @HaoChiBao)
-
Define the API for in-page Agents to use a site's declared tools webmachinelearning/webmcp#51 (1 comment by @domfarolino)
-
Hosting official TypeScript types for WebMCP webmachinelearning/webmcp#190 (1 comment by @beaufortfrancois)
-
-
17 pull requests opened.
-
Drop the stale native-agents note from the security questionnaire webmachinelearning/webmcp#269, by @lazizbekravshanov.
-
Define input-length mitigations in code points webmachinelearning/webmcp#265, by @HaoChiBao.
-
Clarify observation context management webmachinelearning/webmcp#264, by @1aifanatic.
-
doc: clear sentence regarding agents in frames webmachinelearning/webmcp#260, by @marcianoskate.
-
Drop closed executeTool issue links from explainers webmachinelearning/webmcp#259, by @HaoChiBao.
-
[codex] Add ChatGPT Desktop to implementation status webmachinelearning/webmcp#258, by @nmccormack-oai.
-
Add `outputSchema` to `ModelContextTool` and `RegisteredTool` webmachinelearning/webmcp#254, by @beaufortfrancois.
-
Add `debugging` member to ToolAnnotations webmachinelearning/webmcp#253, by @beaufortfrancois.
-
Change `executeTool()` inputObject type from `object` to `any` webmachinelearning/webmcp#251, by @beaufortfrancois.
-
Fix mismatching tool parameters webmachinelearning/webmcp#250, by @domfarolino.
-
Preserve in-flight executions after unregistration webmachinelearning/webmcp#248, by @domfarolino.
-
Spec `AbortSignal` integration with tool execution webmachinelearning/webmcp#247, by @domfarolino.
-
Accept an object in executeTool() instead of a JSON string webmachinelearning/webmcp#246, by @beaufortfrancois.
-
Add ToolActivatedEvent and ToolCancelEvent interfaces webmachinelearning/webmcp#245, by @beaufortfrancois.
-
Fix normative defects in tool registration, notification, and observation webmachinelearning/webmcp#244, by @emecii.
-
Change `RegisteredTool#inputSchema` type from `DOMString` to `object` webmachinelearning/webmcp#241, by @beaufortfrancois.
-
Validate `exposedTo` before registering `AbortSignal` algorithm webmachinelearning/webmcp#240, by @beaufortfrancois.
-
-
12 pull requests merged.
-
doc: clear sentence regarding agents in frames webmachinelearning/webmcp#260.
-
[codex] Add ChatGPT Desktop to implementation status webmachinelearning/webmcp#258.
-
Drop closed executeTool issue links from explainers webmachinelearning/webmcp#259.
-
Preserve in-flight executions after unregistration webmachinelearning/webmcp#248.
-
Spec `AbortSignal` integration with tool execution webmachinelearning/webmcp#247.
-
Fix mismatching tool parameters webmachinelearning/webmcp#250.
-
Accept an object in executeTool() instead of a JSON string webmachinelearning/webmcp#246.
-
Fix normative defects in tool registration, notification, and observation webmachinelearning/webmcp#244.
-
Change `RegisteredTool#inputSchema` type from `DOMString` to `object` webmachinelearning/webmcp#241.
-
Spec the `executeTool()` API webmachinelearning/webmcp#226.
-
Validate `exposedTo` before registering `AbortSignal` algorithm webmachinelearning/webmcp#240.
-
Add implementation-status.md to track browser implementation status webmachinelearning/webmcp#235.
-
-
1 pull request closed.
-
Add ToolActivatedEvent and ToolCancelEvent interfaces webmachinelearning/webmcp#146.
-
-
11 pull requests received 63 new comments.
-
Spec `AbortSignal` integration with tool execution webmachinelearning/webmcp#247 (15 comments by @beaufortfrancois, @domfarolino)
-
Spec the `executeTool()` API webmachinelearning/webmcp#226 (13 comments by @beaufortfrancois, @domfarolino)
-
Add implementation-status.md to track browser implementation status webmachinelearning/webmcp#235 (12 comments by @beaufortfrancois, @bwalderman, @domfarolino, @Idan-Levin, @victorhuangwq)
-
Validate `exposedTo` before registering `AbortSignal` algorithm webmachinelearning/webmcp#240 (6 comments by @beaufortfrancois, @domfarolino, @markafoltz)
-
[codex] Add ChatGPT Desktop to implementation status webmachinelearning/webmcp#258 (5 comments by @domfarolino, @nmccormack-oai, @victorhuangwq, @w3cbot)
-
Change `RegisteredTool#inputSchema` type from `DOMString` to `object` webmachinelearning/webmcp#241 (5 comments by @beaufortfrancois, @domfarolino)
-
Preserve in-flight executions after unregistration webmachinelearning/webmcp#248 (2 comments by @beaufortfrancois, @domfarolino)
-
Add ToolActivatedEvent and ToolCancelEvent interfaces webmachinelearning/webmcp#146 (2 comments by @beaufortfrancois, @domfarolino)
-
Drop the stale native-agents note from the security questionnaire webmachinelearning/webmcp#269 (1 comment by @w3cbot)
-
webmachinelearning/translation-api
-
7 issues opened.
-
Support for batched translation webmachinelearning/translation-api#85, by @sushraja-msft.
-
What are the **safety/privacy** semantics (sanitization, the "safe mode" idea)? | Needs to be defined webmachinelearning/translation-api#84, by @toreini.
-
What **interoperability guarantees** apply (which markup/attributes are preserved, translated, or may change)? webmachinelearning/translation-api#83, by @toreini.
-
What is the input **representation**, an opaque *string + format/`mimeType` tag*, or a *structured DOM-to-DOM* transform? webmachinelearning/translation-api#82, by @toreini.
-
Do we support input beyond plain text at all? webmachinelearning/translation-api#81, by @toreini.
-
Add support for Romanized languages webmachinelearning/translation-api#79, by @tomayac.
-
Document security considerations for HTML translation webmachinelearning/translation-api#78, by @reillyeon.
-
-
6 issues received 11 new comments.
-
Input formats for translation and language detection webmachinelearning/translation-api#60 (2 comments by @sushraja-msft, @toreini)
-
Support for batched translation webmachinelearning/translation-api#85 (2 comments by @sushraja-msft, @toreini)
-
Document security considerations for HTML translation webmachinelearning/translation-api#78 (2 comments by @anssiko, @toreini)
-
What **interoperability guarantees** apply (which markup/attributes are preserved, translated, or may change)? webmachinelearning/translation-api#83 (2 comments by @toreini)
-
What is the input **representation**, an opaque *string + format/`mimeType` tag*, or a *structured DOM-to-DOM* transform? webmachinelearning/translation-api#82 (2 comments by @toreini)
-
Clarification on Definition of "On-Device Models" Based on TAG Feedback webmachinelearning/translation-api#72 (1 comment by @toreini)
-
-
1 pull request received 3 new comments.
-
Fixed warnings, and made changes to the privacy considerations making it more consistent with the explainer webmachinelearning/translation-api#74 (3 comments by @reillyeon)
-
webmachinelearning/awesome-webmcp
-
1 issue opened.
-
Add TweakTown to Websites webmachinelearning/awesome-webmcp#29, by @camwilmot.
-
-
11 pull requests opened.
-
Add Orbilo to Tools webmachinelearning/awesome-webmcp#31, by @maxkruger.
-
Add AIC assurance tooling webmachinelearning/awesome-webmcp#30, by @VPAI-Grok.
-
Add simple-webmcp webmachinelearning/awesome-webmcp#28, by @emingure.
-
Add FlightQueue to Websites webmachinelearning/awesome-webmcp#26, by @jackculpan.
-
Add Airport Lounge List to Websites webmachinelearning/awesome-webmcp#25, by @jackculpan.
-
Add WindTunnel and WebMCP Kit webmachinelearning/awesome-webmcp#24, by @Idan-Levin.
-
Add WebMCP Today to Tools webmachinelearning/awesome-webmcp#23, by @robertn702.
-
Add WebMCP Attack Surface Scanner to README webmachinelearning/awesome-webmcp#22, by @ElonMusk2002.
-
Add QR Code Crafter to Demos webmachinelearning/awesome-webmcp#21, by @brandaspect.
-
Add webmcp-go to Frameworks webmachinelearning/awesome-webmcp#20, by @seunghan91.
-
Add webmcp-django to Frameworks webmachinelearning/awesome-webmcp#19, by @seunghan91.
-
-
2 pull requests merged.
-
Add webmcpify to Tools webmachinelearning/awesome-webmcp#17.
-
Add Stacktree (Websites) and agentk (Frameworks) webmachinelearning/awesome-webmcp#15.
-
-
1 pull request received 1 new comment.
webmachinelearning/prompt-api
-
2 issues opened.
-
Spam webmachinelearning/prompt-api#216, by @jo8296452-art.
-
Allow clone() to add/override initialPrompts webmachinelearning/prompt-api#213, by @ibnesayeed.
-
-
1 issue closed.
-
Spam webmachinelearning/prompt-api#216.
-
-
3 issues received 5 new comments.
-
Proposal: Introduce Categorical Sampling Modes for the Prompt API webmachinelearning/prompt-api#203 (2 comments by @anssiko, @isaacahouma)
-
Extend the Message format to support tool calling, extended thinking, and multimodality webmachinelearning/prompt-api#209 (2 comments by @anssiko, @jingyun19)
-
-
1 pull request opened.
-
Remove samplingMode default from IDL; update comments webmachinelearning/prompt-api#215, by @michaelwasserman.
-
-
1 pull request merged.
-
Remove samplingMode default from IDL; update comments webmachinelearning/prompt-api#215.
-
-
1 pull request closed.
-
Add initial devcontainer configuration webmachinelearning/prompt-api#167.
-
-
1 pull request received 2 new comments.
-
tool use without auto execution webmachinelearning/prompt-api#162 (2 comments by @anssiko, @jingyun19)
-
webmachinelearning/awesome-built-in-ai
-
3 issues opened.
-
[Add Resource] Semantic Sticky Board: Client-Side Semantic Clustering with Semantic Embedder API webmachinelearning/awesome-built-in-ai#5, by @KenjiBaheux.
-
[Add Resource] Rakuten: Client-Side Observability & OpenTelemetry for Prompt API webmachinelearning/awesome-built-in-ai#4, by @KenjiBaheux.
-
Consider adding AAC Board webmachinelearning/awesome-built-in-ai#3, by @tomayac.
-
-
1 issue received 1 new comment.
webmachinelearning/webmcp-types
-
1 issue opened.
-
Automatically infer args from input schema? webmachinelearning/webmcp-types#5, by @MiguelsPizza.
-
-
1 issue received 1 new comment.
-
Automatically infer args from input schema? webmachinelearning/webmcp-types#5 (1 comment by @beaufortfrancois)
-
-
4 pull requests opened.
-
Infer execute input from inputSchema webmachinelearning/webmcp-types#6, by @MiguelsPizza.
-
add ToolExecuteCallbackOptions to ToolExecuteCallback webmachinelearning/webmcp-types#4, by @beaufortfrancois.
-
Add executeTool() method webmachinelearning/webmcp-types#3, by @beaufortfrancois.
-
Update RegisteredTool#inputSchema type from string to object webmachinelearning/webmcp-types#2, by @beaufortfrancois.
-
-
2 pull requests merged.
-
add ToolExecuteCallbackOptions to ToolExecuteCallback webmachinelearning/webmcp-types#4.
-
Update RegisteredTool#inputSchema type from string to object webmachinelearning/webmcp-types#2.
-
-
1 pull request reopened.
-
Add executeTool() method webmachinelearning/webmcp-types#3, by @beaufortfrancois.
-
-
1 pull request received 3 new comments.
webmachinelearning/writing-assistance-apis
-
1 pull request received 1 new comment.
-
Add constraints and future outlook for speed preference webmachinelearning/writing-assistance-apis#103 (1 comment by @anssiko)
-
webmachinelearning/proposals
-
1 issue received 1 new comment.
webmachinelearning/webnn-samples-test-framework
-
5 pull requests opened.
-
Fix timeout multiplier mutating shared config webmachinelearning/webnn-samples-test-framework#67, by @zhibo1-wang.
-
Add browser close timeout webmachinelearning/webnn-samples-test-framework#66, by @zhibo1-wang.
-
Refactor report table webmachinelearning/webnn-samples-test-framework#65, by @zhibo1-wang.
-
Add wildcard filter support webmachinelearning/webnn-samples-test-framework#64, by @zhibo1-wang.
-
Split util.js webmachinelearning/webnn-samples-test-framework#63, by @zhibo1-wang.
-
-
5 pull requests merged.
-
Fix timeout multiplier mutating shared config webmachinelearning/webnn-samples-test-framework#67.
-
Add browser close timeout webmachinelearning/webnn-samples-test-framework#66.
-
Refactor report table webmachinelearning/webnn-samples-test-framework#65.
-
Add wildcard filter support webmachinelearning/webnn-samples-test-framework#64.
-
Split util.js webmachinelearning/webnn-samples-test-framework#63.
-
-
5 pull requests received 5 new comments.
-
Fix timeout multiplier mutating shared config webmachinelearning/webnn-samples-test-framework#67 (1 comment by @ibelem)
-
Add wildcard filter support webmachinelearning/webnn-samples-test-framework#64 (1 comment by @ibelem)
-