Files
next-ai-draw-io/lib/access-code.ts
T
Dayuan Jiangandgithub-actions[bot] 0b63e28e5a feat: MCP feature parity, the web app on the MCP core, and provider fixes (#951)
* fix(chat): close credential leaks and harden the chat route

- Vertex: a client-supplied base URL only works with the client's own Vertex key
- Accept only data: URLs for file parts in every message, so the server never downloads them
- Output budget retry accounts for the thinking budget Bedrock/Anthropic add, and reads
  Volcengine, DashScope, SGLang and vLLM rejections; falls back to 16000 once
- x-max-output-tokens can only lower the budget on server credentials
- On server credentials only server models or AI_MODEL entries can be used
- Drop tool results together with the invalid tool calls they belong to
- Count quota tokens as input + output (cached tokens were counted twice)
- Private-URL check for custom base URLs, end Langfuse traces on error/abort/early return
- Fix repairToolCall ordering and placeholder, align edit_diagram prompt with operations
- Panel Bedrock keys are read from ADMIN_AWS_*; forward the access code to EdgeOne
- isMinimalDiagram only treats root cells as an empty canvas

* fix(api): require access codes and limit sizes on helper routes

- Shared checkAccessCode for validate-diagram, validate-model, parse-url, verify-access-code
- parse-url: 5 MB streamed body limit; validate-diagram: 5 MB image limit
- validate-model refuses redirects when private URLs are blocked
- Admin settings state shared across module instances via globalThis
- Server model ids: unique slugs (non-ASCII names encoded), duplicates rejected
- Panel Bedrock credentials stored as ADMIN_AWS_* so the DynamoDB client keeps its own
- Locale redirect keeps basePath and query; EdgeOne function drops open CORS and checks the access code
- Providers payload reports whether .env sets a default model

* fix(chat): keep saved diagrams and pages when restoring, editing and retrying

- Restored sessions no longer replay the last display_diagram over the saved diagram
- Failed or stopped edit_diagram restores the canvas
- Message snapshots keep the full multi-page document
- "Improve with suggestions" uses the normal send path (headers, xml, retry counters)
- Editing a message keeps its file/URL sections; cached example edits work
- New chat's first autosave no longer resets the UI
- Validation retries counted per user turn; validate-diagram sends the access code
- Cached examples only match the example files on an empty canvas
- Template sends keep attachments and wait for extraction

* fix(diagram): fix autosave staleness and XML repair corrupting valid diagrams

- Autosave guard reads refs, so edits after a theme or dark mode switch are kept
- Duplicate-id check and rename run per page; repair loop no longer quadratic
- autoFixXml no longer breaks style values, rich text " or single-line cells
- extractCompleteMxCells keeps the cell after a self-closing cell
- Better truncation detection; object/UserObject wrapped cells are editable
- Exports for thumbnail, PNG and save are routed by tag instead of a shared resolver
- History stores the full document; storage errors are reported, no auto-deletion of chats
- IndexedDB connection reopens after errors; focus refresh throttled
- Keep ?session= on locale redirect, map zh-Hant to zh-tw for draw.io

* fix(chat-input): stop template dialogs from sending and fix attachment races

- Template dialogs no longer submit the outer chat form
- Sending is blocked while files or URLs are still extracting
- File and URL extraction no longer drop or resurrect entries
- IME composition Enter no longer sends
- Tool call cards show the error text; keyboard handling on cards fixed
- Template import available when empty, edit dialog resets, saved templates refresh
- Only png/jpeg/gif/webp images accepted, SVG sent as text; PDF objects released
- parse-url request sends the access code

* fix(model-config): keep model selection valid and fix admin panel edge cases

- Fall back to the default server model when a saved one disappears
- Sync model config across tabs
- Validation uses the base path and sends the access code
- Model ids edited as drafts (no empty, duplicate or padded ids)
- Credential changes reset validation; stale validation results are dropped
- Admin: generateId over HTTP, env-locked group switches, discard and toggle fixes,
  clearing a secret field keeps the saved key, first provider not auto-default when .env sets AI_MODEL
- Model selector items use unique values

* fix(electron): decrypt keys after ready and harden navigation and IPC

- Apply preset env after app ready, so Windows/Linux get decrypted keys
- Never re-encrypt ciphertext; restore env when switching or removing presets
- Block navigation away from the app, open external links in the browser, check IPC senders
- Keep inherited proxy settings, default NO_PROXY for localhost
- Serialize server start/restart, kill stuck processes, follow port changes
- Atomic config writes, keep corrupt files as backups, remember the server port
- Menu and settings window stay in sync; dev script gets the decrypted preset env
- Use app.isPackaged, parse inline .env comments, drop .env files from the bundle

* fix(mcp-server): fix XSS and crashes, make XML validation strict

- Validate and escape the mcp session id; only serve localhost Host/Origin
- Malformed URLs and session ids return errors instead of crashing the process
- Strict XML syntax check with saxes (linkedom never reports parse errors)
- autoFixXml no longer corrupts valid XML; attribute newlines serialized as entities
- Sessions stay alive while polled; browser pushes carry a base version (409 on conflict)
- Page tools respect the edit gate; UTF-8 bodies decoded correctly
- Export replies matched to requests and serialized; xml sync export handled
- UserObject/object cells addressable by id; history restored by stable id; logs off stdout

* fix(mcp-server): make edit_diagram all-or-nothing and fix preview sync races

- edit_diagram applies nothing when any operation fails, rejects invalid or
  multi-cell new_xml, validates only the target page, and returns the
  current page XML on every rejection (including stale edits)
- Fix get_diagram reading the old diagram right after an AI write: the
  preview pushed its sync reply with a newer version than it was taken at
- Keep a user edit that loses the race with an AI write in history and
  tell the user in the preview
- Autofix removes only exact foreign tags (a stray <mxGraph/> deleted
  <mxGraphModel>), fixes tag case, drops orphan <mxPoint>s, and rejects
  unknown element names in model XML
- Edit empty and compressed pages; PNG exports use the page on screen;
  tag download exports; reload from the server after a page export
- Expand ~ in paths, tell the model when the browser sync timed out,
  use registerPrompt, require SDK ^1.31.0

* feat(mcp-server): bring the web app's drawing knowledge to MCP

- Add a drawing guide adapted from the web system prompt (layout, edge
  routing, styles, minimal style, editing rules), returned by
  start_session, a new get_drawing_guide tool and the diagram-workflow prompt
- Add get_shape_library with the 30 icon libraries; the build copies
  docs/shape-libraries into dist and CI checks the packed files
- Accept bare mxCell lists in create_new_diagram and add_page; the server
  adds the wrapper and root cells
- Send server instructions, shorten create_new_diagram's description to
  fit Claude Code's 2,048 character limit, and annotate every tool
- Fix dead links and the totals in docs/shape-libraries/README.md

* feat(mcp-server): add screenshot_diagram so the model can check its render

- New read-only screenshot_diagram tool returns the rendered page as a PNG
  plus the web app's visual checklist (overlaps, edges crossing shapes,
  readability, layout, rendering errors), replacing the web app's
  separate vision model with the host model's own vision
- PNG exports use draw.io's width and pageId options: screenshots stay
  under ~140,000 base64 characters and page exports no longer swap the
  page on screen
- Fail fast with a clear message when the preview tab stopped polling
  (browsers throttle background tabs)
- Mention the screenshot step in the drawing guide and instructions

* feat(mcp-server): auto-save each session's diagram to a .drawio file

- Save the latest diagram of every session 1 second after each change
  (AI write, browser edit, history restore) to ~/.next-ai-drawio/<id>.drawio,
  keep the newest 50, flush on shutdown; DRAWIO_DATA_DIR changes the folder
  and "off" disables it, like the web app's IndexedDB sessions
- start_session names the file, so a resumed conversation can reopen the
  diagram with load_diagram after the MCP process restarted
- Fix PNG/SVG exports randomly timing out: a previous export's 10 second
  timer cleared the export in progress, and a late reply could be taken
  for the current one; exports are now numbered

* refactor(mcp-server): move the preview page into src/preview

The 580-line page template in http-server.ts becomes index.html,
preview.css and preview.js, copied to dist/preview by the build and
filled at request time. The rendered page is unchanged apart from the
session id and draw.io origin now coming from a small config script.
Biome skips the folder because of the {{placeholders}}, as it never
linted the old template string either.

* feat(mcp-server): add theme menu, dark mode and editable SVG to the preview

- Pass themes=1 and dark=auto to the draw.io iframe, so the Extras menu
  offers the Theme submenu and draw.io follows the system dark mode; the
  header and dialogs follow it too through prefers-color-scheme
- Download dialog: new Editable SVG (.drawio.svg) format through draw.io's
  xmlsvg export, default name diagram-YYYY-MM-DD, Enter saves and Escape
  closes it and the history dialog
- export_diagram accepts format "drawio.svg" and detects the .drawio.svg
  extension
- Use system fonts instead of Google Fonts, so the page works offline with
  a self-hosted draw.io (DRAWIO_BASE_URL)

* docs(mcp-server): describe the new MCP features in all READMEs

- MCP server README: 13 tools, drawing rules and shape libraries,
  screenshots, all-or-nothing edits, auto-save and how to continue a
  diagram later, DRAWIO_DATA_DIR and DEBUG, offline use with a local
  draw.io, and what to do when a background tab makes exports time out
- Fix two errors: History is a button at the top right of the preview
  page, and exports are not limited to .drawio
- Claude Code plugin README: same tool list, formats and settings
- Root READMEs (English, Chinese, Japanese): short list of what the MCP
  server can do

* fix(mcp-server): fix duplicate page exports and auto-save deleting user files

- Preview page: keep an MCP export open until the server has its result.
  A poll answered before that still saw the request and started the same
  export again, so a parallel page export could write the previous
  page's image into its file
- Auto-save only removes its own mcp-*.drawio files, so a DRAWIO_DATA_DIR
  that also holds the user's diagrams keeps them
- screenshot_diagram captures a page that has no id attribute by loading
  just that page, like export_diagram
- An empty <Array as="points"/> no longer hides orphan mxPoints that
  come after it
- POST /api/state refuses a push without xml, which used to wipe the
  stored diagram
- Clear exportOptions when an export ends, reuse hasCells for the empty
  diagram check, and reword two log lines

* fix: log Editable SVG saves, use local dates in file names, drop unused Electron dialogs

- /api/log-save accepted only drawio, png and svg, so saving as Editable
  SVG (xmlsvg) got a 400 and was never recorded in Langfuse
- The default download name and the template export name used the UTC
  date, which is the previous day on mornings in East Asia
- Remove the Electron openFile/saveFile IPC handlers, their preload
  bindings and types; nothing in the app calls them

* refactor(mcp-server): make the diagram modules usable from the web app

The web app will reuse the MCP server's XML engine instead of its own
copy in lib/utils.ts, so these modules now run in the browser too.

- Relative imports end in .ts, rewritten to .js by tsc
  (rewriteRelativeImportExtensions); Next.js resolves them directly
- Every module uses the global DOMParser/XMLSerializer: native in the
  browser, linkedom in Node via installDomPolyfill. pages.ts parsed with
  linkedom but serialized with the global serializer, which throws in
  the browser
- The saxes syntax check moves to xml-syntax.ts, so the browser does not
  pull in linkedom; it now also rejects undeclared prefixes such as
  xlink:, as the browser does
- Page decompression uses pako and atob instead of node:zlib and Buffer
- hasCells moves to pages.ts, away from the file system code
- The duplicate cell id check counts UserObject/object ids
- wrapCellsInModel drops comments and text before the first cell, which
  the web app accepts today
- validateAndFixXml takes { strict: false } for diagrams with user content
- Web tests run these modules with a browser DOM (jsdom)
- saxes becomes a direct dependency of the web app

* refactor(web): validate and repair diagram XML with the MCP server's engine

- Delete the web app's own copy of the XML checks and repairs from
  lib/utils.ts (1,074 lines). loadDiagram now uses the MCP server's
  validateAndFixXml without the strict checks, because the XML may hold
  the user's own diagram
- display_diagram and append_diagram prepare the model's XML with the new
  shared prepareNewDiagram, also used by the MCP create_new_diagram: wrap,
  validate strictly and auto-fix while it is still a bare model (where
  duplicate ids are renamed), then turn it into an mxfile
- The streaming preview of display_diagram no longer redraws the model's
  raw cells after the tool handler loaded the checked diagram, and drops
  a queued preview once the input is complete. That redraw lost
  auto-fixes and UserObject/object wrappers, so a linked cell lost its
  label; it also showed a second error toast
- The web repair regression tests now run against the MCP functions
- New e2e test checks the canvas content after display_diagram
- Fix the e2e upload tests, whose file input locator also matched the
  template import input

* refactor(web): edit and wrap diagrams with the MCP server's code

- edit_diagram runs the MCP server's editDiagram: every new_xml is checked
  first, one cell per operation, and after the edit only the target page
  is checked, rejecting only errors this edit introduced. An unrelated
  problem elsewhere in the document no longer blocks every edit. The
  error lists each failed operation
- The streaming edit preview uses the MCP applyDiagramOperations
- Delete applyDiagramOperations (292 lines) and wrapWithMxFile from
  lib/utils.ts, and the unused hand-copied scripts/test-diagram-operations.mjs
- One blank document (BLANK_MXFILE) for the web app and the MCP preview,
  replacing four copies
- Saving a .drawio wraps a bare model with normalizeToMxfile
- The empty-diagram check uses hasCells, which also counts cells wrapped
  in a UserObject/object
- DiagramOperation is the MCP type
- The wrapped-cell and empty-diagram tests now run against the MCP code
- New e2e test: edit_diagram changes the canvas, and a failing edit
  leaves it as it was

* refactor(web): share prompt examples and the shape library with the MCP server

- The three XML examples (swimlanes, two edges, waypoints) that the web
  system prompt, the display_diagram description and the MCP drawing
  guide each had a copy of now live in packages/mcp-server/src/xml-examples.ts
- The shape library group list and reader come from the MCP
  shape-library.ts; getShapeLibrary takes the folder, and the web route
  passes docs/shape-libraries under the working directory as before.
  Only the 30 known library names are read. Error texts differ slightly
- next.config.ts traces docs/shape-libraries/*.md for /api/chat, since
  the read now happens in another module
- Every prompt is byte-for-byte unchanged: the web system prompt for five
  model ids with and without minimal style, and the MCP drawing guide,
  compared before and after; the library list and swimlane example match
  the old tool description text exactly

* fix(providers): update the v6 SDK packages and fix Claude and Gemini settings

- Update ai to 6.0.300 and the @ai-sdk providers to their latest v6-line
  versions. @ai-sdk/anthropic 3.0.47 did not know claude-opus-4-7/4-8
  and capped their output at 32000 tokens; 3.0.127 allows 128000
- Drop the fine-grained-tool-streaming beta header for the Anthropic API:
  the provider now streams tool input per tool (eager_input_streaming)
- Claude 4.7 and later reject a non-default temperature/top_p/top_k and
  the extended thinking budget with a 400. A middleware retries once
  without them, so TEMPERATURE and *_THINKING_BUDGET_TOKENS no longer
  break those models
- Prompt caching also reaches Claude on the Anthropic API and OpenRouter;
  before, only Bedrock got a cache marker
- GOOGLE_TOP_K and GOOGLE_TOP_P never reached Gemini: they were sent as
  Google provider options, which drops them. They are call settings now.
  GOOGLE_CANDIDATE_COUNT and GOOGLE_REASONING_EFFORT, which the provider
  does not support, are removed
- Add @ai-sdk/openai-compatible as a direct dependency

* refactor: simpler streaming preview and small AI SDK cleanups

- useChat throttles streamed message updates (experimental_throttle,
  150 ms), replacing the two hand-written 150 ms timers of the
  display_diagram and edit_diagram previews (94 lines less). The preview
  now only runs while the input streams; once it is complete the tool
  handler takes over, so a queued preview can no longer redraw an edit
  the handler rejected and rolled back. Measured on a streamed 60-cell
  diagram: 41 redraws at least 97 ms apart, before 37 with gaps down to
  48 ms
- The diagram check endpoint uses streamText with Output.object instead
  of the deprecated streamObject, and returns its fixed result as a plain
  text response; new route test
- Import createGateway/gateway from ai and drop the direct
  @ai-sdk/gateway dependency
- The per-request message structure logs only print with
  DEBUG_LLM_PAYLOAD=true
- Remove an empty onFinish callback

* refactor(chat): check the last tool part with the SDK's isToolUIPart

Drop the hand-written MessagePart and ChatMessage types that only served
this check.

* refactor(providers): one model factory for chat and the settings Test button

- getAIModel resolves credentials (client key, server env vars, the
  existing SSRF rules) and createModel builds the model by SDK. The
  provider-by-provider switch shrinks from 24 cases to the few that
  differ (lib/ai-providers.ts 1531 -> 1106 lines)
- /api/validate-model calls getAIModel instead of its own 24-case switch
  (503 -> 175 lines), which had drifted from the chat: it built Azure
  with createOpenAI, Kimi and MiMo with createOpenAI instead of
  createDeepSeek, and the official OpenAI endpoint with Chat Completions.
  A passing test now means the chat works
- Plain OpenAI-compatible providers (SiliconFlow, SGLang, ModelScope,
  GLM, Qwen, Qiniu, Novita, Atlas Cloud, EdgeOne, Doubao, MiniMax in
  OpenAI mode, AIHubMix on a custom URL) use @ai-sdk/openai-compatible,
  which reads reasoning_content, so their reasoning shows, and accepts
  SGLang's stream as is (its 95-line stream rewrite is gone).
  includeUsage keeps token usage for quotas. <think> tags in their text
  become reasoning (extractReasoningMiddleware)
- SGLang without a base URL used OpenAI's endpoint; it now defaults to
  http://127.0.0.1:8000/v1 like the Test button did
- Chat requests to a client base URL refuse redirects, as the Test
  button already did (redirectGuardedFetch moves to lib/ssrf-protection)
- The Test button streams like the chat (the ModelScope special case is
  gone), times out after 15 s, does not retry, asks the model to call a
  ping tool and warns when it answers without one, and tests all models
  at once. The time each test took shows on its check mark
- Unknown provider names are rejected with Object.hasOwn, and the error
  texts list providers from PROVIDER_INFO instead of hand-kept lists

* feat(settings): link to each provider's key page and clean up base URLs

- A "Get API key" link next to the API Key field for the 19 providers
  that have a key page (from env.example and the providers' docs). 17
  answered 200 to curl; OpenAI's is behind a Cloudflare challenge and
  DeepSeek's behind a regional block, both checked in Chrome
- Base URLs drop spaces, trailing slashes and a pasted endpoint path
  (/chat/completions, /completions, /messages, /responses), which the
  SDK would otherwise append a second time and get a 404. getAIModel does
  this for the chat and the Test button; the field does it on blur and
  shows the URL requests go to

* feat(errors): classify provider errors and show a hint the user can act on

- lib/llm-errors.ts sorts an error into about a dozen kinds (key
  rejected, no access, unknown model, no credit, rate limited, context
  too long, no image input, no tool calls, output cut off, provider down,
  cannot connect, timeout): first texts that name the cause precisely,
  then the HTTP status code, then general texts. It unwraps RetryError and
  hides keys and Bearer tokens in the provider's message
- The chat route uses it for errors before the stream and, through
  toUIMessageStreamResponse's onError, for errors in the stream. Errors
  of the model's own tool call stay as they are: the same text goes back
  to the model so it can fix the call
- The chat shows the hint in the user's language, then the provider's
  message; a rejected key, missing access or unknown model adds an "Open
  model settings" button. The Test button shows the same hints
- Fixes: our message "API key is required when using a custom base URL"
  was replaced by "Authentication failed" because it contains "key"; a
  provider's "Rate limit exceeded" opened this site's quota toast; an
  error body like {"error": ...} was shown as raw JSON; the Test button
  matched "401" in the message, where providers rarely put it
- Remove the string matching fallbacks in the chat panel

* refactor(settings): drop an unused index parameter

* feat(settings): fetch the model list from any provider and flag models that cannot draw

The "Fetch models" button asks the provider for its models (OpenAI-style
/models, Anthropic, Google, Ollama, OpenRouter, Vercel Gateway, AIHubMix)
and shows them in a searchable picker. This replaces the route that only
worked for AIHubMix.

A snapshot of models.dev (MIT) says which models support tool calls.
Models without them get a "no tool calls" badge in the picker and a hint
in the model list, since drawing needs tool calls. Refresh the snapshot
with scripts/update-model-catalog.mjs.

* fix(mcp-server): reject text between tags, which draw.io cannot open

draw.io reads any text inside a page as compressed page data, so a stray
text node makes the whole page fail with an atob error. gpt-5-mini sends
new cells with a literal "\n" between the tags; the edit card said
Complete while draw.io showed the error and kept the old diagram.

Validation now reports text between tags, and auto-fix turns a literal
\n, \t or \r between tags into whitespace. Other text goes back to the
model as an error. The compressed data directly under <diagram> is fine.

* fix(chat): clearer provider errors and no empty bubble, found with real models

- An error object sent inside the stream (OpenRouter's { code, message })
  showed as "[object Object]"; its message and status code are read now.
- A problem+json "detail" is added to the message: NVIDIA only said
  "Gone" for a retired model. 410 counts as model not found.
- "Cannot connect to API" from the SDK gets the connection hint.
- Text that is only whitespace (Kimi K2.6 sends a space before a tool
  call) no longer shows an empty bubble.
- allowSystemInMessages stops the warning on every request. Our system
  messages carry cache points; a client's own system messages are already
  dropped by the empty-content filter.

* feat(providers): suggest the Claude 5 and GPT 6 models, show GPT 6 reasoning

The suggested models stopped at Claude Opus 4.8 and GPT 5.5. They now
start with Claude Opus 5.5, Sonnet 5.5 and Fable 5.1, and gpt-6.1-sol,
gpt-6-sol, gpt-6-luna and gpt-6-astra (ids checked against the provider
lists). The Bedrock list is unchanged until its ids are checked.

The reasoning summary was only turned on for model ids containing o1,
o3, o4 or gpt-5, so GPT 6 models showed no thinking. It now matches the
o-series and gpt-5 or later by version, like the OpenAI SDK.

* fix(providers): keep thinking on the newest Claude models, fix the Bedrock ids

Measured on Bedrock: Claude Opus 4.7, 4.8 and every Claude 5 model reject
a thinking budget ("thinking.type.enabled") and a temperature, and all of
them accept adaptive thinking. The retry used to drop thinking, so with
a thinking budget set these models did not think at all. It now switches
to adaptive thinking with display "summarized"; without that setting the
models think but send no thinking text to show.

The suggested Bedrock ids had no region prefix, and the newer models only
answer through an inference profile id: "anthropic.claude-sonnet-5-5"
fails with "on-demand throughput isn't supported". Each suggested id was
called once; the Claude ones now start with "global.", Llama, Pixtral and
Nova 2 Lite with "us.". Nova Premier (end of life) and the ids without a
working profile are gone. The docs example had the same problem and a
wrong date.

* fix(chat): an edit after a broken edit call no longer fails, found with Opus 5.5

- Claude Opus 5.5 sent an edit with invalid JSON, then the same edit
  again. The first call's streamed preview was never undone: its input
  has no operations, and the undo sat behind that check. The second edit
  then started from the preview, failed on a duplicate id, and the model
  had to try a third time. The undo now runs first, and an edit that
  starts in the same render uses the undone diagram.
- The SDK passes an invalid tool call's error as a string, which was
  wrapped as a provider error. streamErrorText keeps it as the text the
  model reads.
- Bedrock's "on-demand throughput isn't supported" gets the model id hint.
- The thinking header uses the page language ("Thought for 1 second" in
  English), from the dictionary entries that were already there.

* chore(mcp-server): release the new features as 0.3.0

npm has 0.2.3. The earlier commits raised the version step by step to
0.6.0, but they ship together as one release.

* style: auto-format with Biome

* fix(chat): draw the built-in examples again and undo edit previews on errors

Found by the PR review:
- The built-in examples showed a finished card and an empty canvas. They
  are answered in the browser, never reach the tool handler, and relied
  on the final redraw that an earlier commit removed. The example branch
  now loads its diagram itself.
- When the request failed while an edit was streaming (a provider error,
  a lost connection), its preview stayed on the canvas. The error handler
  now restores the diagram from before the preview.
- The model picker could not scroll with the wheel or touch: the settings
  dialog blocks those events outside itself, and the picker is rendered
  outside it. The popover is modal now.
- A fetch error and the open picker stayed when switching providers.
- Editing a model id kept the old test warning and response time, which
  also hid the "may not be able to draw" hint for the new id.

* fix(mcp-server): keep both pages when get_diagram meets a page export, and more review fixes

Found by the PR review, each with a test that failed first:
- get_diagram during a page export returned the one-page projection on
  screen as the whole document (6 of 6 times when timed so). The preview
  page no longer answers a sync while a projection shows, and syncs after
  reloading, so the poll that restores the real document exports it.
- Exports are numbered on the server too: a late result of an export that
  timed out was saved as the next export's file.
- In Chrome, a new_xml with a syntax error counted the <parsererror>
  element as a second cell, so the web app rejected edits that auto-fix
  repairs ("must contain exactly one cell").
- hasCells missed single-quoted ids, so screenshot_diagram called such a
  diagram empty and auto-save never created its file.
- A literal \n directly under a <diagram> that has a model passed
  validation; only text-only pages are compressed data.
- A wrapped mxCell repeating its UserObject's id took the wrapper's place
  in edits, so delete and update left an empty or nested wrapper.
- Bare cells with a shape or edge id of "0" or "1" are rejected with a
  clear message instead of being renamed, which broke their edges.
- DRAWIO_DATA_DIR expands ~, which JSON configs pass on as it is.

* fix(server): count quota by the key actually used, and more review fixes

Found by the PR review, each with a test that failed first:
- Quota: any key header skipped it, even one the provider never reads
  (x-aws-access-key-id with OpenAI), so a request ran on the server's
  key without being counted. The check now runs after the model is
  resolved and uses usesServerCredentials. On main already.
- usesServerCredentials read the raw base URL; "/" cleans up to none, so
  an Ollama request ran on the server's key past the server-model check.
- SGLang's default 127.0.0.1:8000 only fills the settings form. Chat and
  the model list used it as a real address, so the server called its own
  machine even with private URLs blocked. Now a base URL is required.
- With a user's OpenAI key and no base URL, the SDK read the server's
  OPENAI_BASE_URL. The official endpoint is now passed. On main already.
- The Test button refused nothing on the server's keys (Ollama Cloud),
  and a 15 s timeout reported "connected, no tool call".
- The model list for Ollama without a base URL came from ollama.com while
  chat went to the server's Ollama.
- Bedrock's "Too many tokens, please wait" counted as context too long.
- On the server's keys the provider's error text stays in the server log;
  it can name the server's AWS account, role or internal hosts.
- Desktop app: the preset keys are the user's own (NEXT_AI_DRAWIO_DESKTOP),
  so Max Output Tokens can be raised and keyless models in settings work
  again. A launch that found the remembered port taken no longer replaces
  it, which hid the user's chats and settings for good.

* ci: run auto-format with the Biome version in package.json again

package.json moved to Biome 2.5.7 but the auto-format job stayed on
2.4.13. The two format some files differently, so on this PR the bot
reformatted tests/unit/log-save-route.test.ts with 2.4.13 and the lint
job, on 2.5.7, then failed on it. The pin now matches package.json, as
its comment asks, and the file is back in 2.5.7's format.

* fix(chat): keep the canvas after an unrelated error, and more review fixes

Found by the second PR review:
- After a streamed edit, an older render of the stream stored the edit's
  original diagram again, and the next failed request (no quota, a lost
  connection) put that old diagram back on the canvas. The tool handler
  now marks its call as handled, so the preview code leaves it alone.
- An edit applied before the UI showed an earlier broken edit's error was
  erased when that error undid its preview, or was built on that preview.
  The handler now starts from the diagram before all unhandled previews,
  and reads the diagram state that updates at once.
- A failed or stopped display_diagram left its half drawn diagram on the
  canvas. Its preview is undone now, like an edit's.
- "New chat" cleared a chat that could not be saved (storage full).
- The settings dialog showed a model list, a fetch error or a test result
  on the provider that was opened after the request started, and marked a
  model id changed during the test as tested.
- A tool call with broken JSON was shown as cut off by the output limit.
- History entries and session thumbnails could pair with a later diagram
  when draw.io answered an export late.
- A server model saved before non-ASCII provider names got into the id
  was reset to the default model.

* fix(mcp-server): count a one-page view only for that page, and more review fixes

Found by the second PR review:
- get_diagram with a page selector, or a rejected edit's error, counted
  the whole document as seen, so an edit on another page could overwrite
  the user's change there. A one-page view now counts for all pages only
  if the others are unchanged; otherwise the reply says to get them.
- add_page accepted shapes with the root cell ids "0" and "1" and renamed
  them, breaking their edges. The check also missed ids on UserObject
  wrappers and ids written with spaces around the "=".
- Root cells written over two lines were kept as an extra layer, cells
  with id = "a" did not count as cells, and CDATA text before a page's
  model passed the check although draw.io cannot open the page.
- Auto-save cleanup deleted the user's own files that start with mcp-.
  Only names in the session id format are removed now.
- Restoring a history entry dropped edits made in the browser since the
  last entry. They are added to history first.
- A session whose state expired showed a blank page, and the next change
  overwrote its auto-save file. The saved file is loaded instead.
- An edit on a page export's one-page projection, made before the real
  document was back, replaced the whole document.
- A late sync reply could overwrite a newer edit: each sync export is
  numbered, and the server ignores replies older than the current state.
- screenshot_diagram could return another session's image after
  start_session ran during its retries.

* fix(server): use the keys the user sent, and more review fixes

Found by the second PR review:
- With AWS_BEARER_TOKEN_BEDROCK set on the server, a request with the
  user's AWS keys ran on the server's token: the Bedrock SDK prefers it.
  Checked with Bedrock: invalid user keys used to get an answer.
- An OpenAI key with the official URL filled in (the settings form does
  that) went to the Responses API. Back to main's rule: a configured base
  URL uses Chat Completions.
- A user's Ollama key went to the server's OLLAMA_BASE_URL, for chat and
  for the model list. Like every other provider, it goes to the user's
  base URL or Ollama Cloud.
- The server's keyless Ollama and EdgeOne were not counted in the quota.
- AI_MODEL models on the server's keys ran on any provider with a server
  key, not only on AI_PROVIDER.
- A user's Azure key without a base URL used the server's resource name.
- The admin panel's Test button failed whenever access codes were set.
- DeepSeek's errors in the stream (plain text) were shown as they were,
  without a hint and also on the server's keys. Bedrock's throttling in
  the stream was not recognised as a rate limit.
- The EdgeOne function accepted text/plain; x=application/json, which
  other sites can send without a CORS preflight.
- Desktop app: a launch that found the old port taken for a moment (the
  previous version still quitting after an update) remembered the new
  port for good. The new port is kept only when Windows reserves the old
  one. A failed read of the presets file moved it aside as corrupt, and a
  save could then replace the presets. Switching presets on the same port
  now reloads the page. The dev launcher no longer misses a preset change
  made before or during a restart.

* fix(chat): undo rejected or stopped previews, and more fixes from the third review

- A call the server runs (get_shape_library) still reaches the browser's
  tool handler, and it dropped the stored diagram of an earlier broken
  edit, whose preview then stayed. Only the tools that draw take it now.
- A display_diagram whose final XML fails the checks loads the diagram
  from before its preview again, as a failed edit does.
- After Stop, a tool result that arrives later (a screenshot check still
  running) no longer sends a new request; Stop also skips calls the tool
  handler already took.
- New chat and opening another chat kept nothing of a diagram drawn
  without messages when it could not be saved; now they stay on it.
- The settings dialog drops a model list or test result whose provider
  credentials changed meanwhile, also in another tab.
- A saved provider this version does not know crashed the whole page on
  load; it is skipped.
- The input emptied a moment after the message showed in the chat, so it
  briefly appeared twice (seen as a flaky e2e test).
- The desktop app's preset switch on the same port refetches the server
  models instead of reloading the page, which lost unsent attachments.

* fix(mcp-server): recover sessions in one place, and more fixes from the third review

- A session whose state expired was recovered from its auto-save file only
  for the preview page; the tools built on their older copy and then
  overwrote the file. They now recover it first (restoreSavedSession).
- A preview tab that missed the last AI write pushed its older diagram
  over the recovered one after a restart. It now shows the recovered
  diagram and keeps its own copy in History.
- An empty record of what the model has seen (after load_diagram or a
  page tool on unseen changes) no longer lets one page of a multi-page
  document count for all, and get_diagram counts a page only once found.
- History: a thumbnail goes only to the entry it shows, the cached image
  never belongs to an older diagram, a re-serialized copy adds no entry,
  and a cleared document with its own pages is kept before a restore.
- The root cell id check reads attributes one by one: rack-id="1" or id
  text inside a label no longer counts.
- A compressed page counts as having cells; a saved file that could not
  be read is never written over.

* fix(server): keep users' keys at their own endpoints, and more fixes from the third review

- Bedrock: a user's AWS keys no longer go to an endpoint the server sets
  in AWS_ENDPOINT_URL_BEDROCK_RUNTIME / AWS_ENDPOINT_URL (read by the
  upgraded SDK), and admin panel keys win over AWS_BEARER_TOKEN_BEDROCK,
  as the Test button checks them. Checked with Bedrock.
- Ollama: a server key without a base URL (admin panel, OLLAMA_API_KEY)
  goes to Ollama Cloud, as env.example says, instead of 127.0.0.1.
- Quota: EdgeOne counts whatever key header comes along, keyless Ollama at
  a private address counts, and their provider texts stay in the log.
- An EdgeOne server model (admin panel, ai-models.json) works: the route
  checked the raw provider header, which holds the name's slug.
- parse-url ends downloads it does not read (too large, PDF, errors).
- Desktop app: the port follows where the chats are (IndexedDB per
  origin) instead of a remembered port, which could hide them for good;
  a same-port restart tells the page to refetch the server models; a
  failed preset switch no longer undoes a newer choice; a presets file
  removed after a failed read can be saved again; .env values quoted from
  start to end keep their inner quotes, as dotenv reads them.

* fix(security): check request sources, regions and endpoints

- Bedrock: a request's AWS region must be a region name. It becomes part
  of the endpoint's host name, so a value such as
  "us-east-1.attacker.example/" sent the server's bearer token or signed
  request to another host.
- MCP preview server: only the preview page itself (Origin equal to the
  Host) or a non-browser client may call it; a page on another localhost
  port could replace the diagram with a plain text POST. History builds
  its thumbnails element by element and shows only SVG data images, so a
  stored value can no longer run script in the preview.
- chat, validate-model, validate-diagram, provider-models and parse-url
  take JSON bodies only, so another website cannot make the user's own
  server (the desktop app, a local install) run models with their keys;
  the desktop app also refuses a foreign Host (DNS rebinding).
- The model list reads at most 2 MB, also through the Gateway SDK, and
  answers only with its own error texts: the URL is the caller's and may
  be an internal address.
- An admin panel provider with its own key and no URL no longer inherits
  the global <P>_BASE_URL, which may be a proxy for another key; OpenAI
  then gets the official endpoint, as its Test. Azure keeps the server's
  resource.

* fix: what the third round broke, and the first batch's review

MCP preview after the server lost a session (it expired, or the MCP
process restarted):
- Every server state has an id, made when the state is created. The tab
  notices a new id even when the version numbers happen to match, and
  every push names the state it was based on, so one based on a lost state
  is refused, also when it comes before the tab's first poll (the server
  recovers the saved file first).
- The tab keeps the newest canvas XML, saved or not. When the server knows
  nothing (no file) or exactly what the tab last saved, the canvas wins and
  is saved, so edits made while the server was down are kept. Otherwise
  the server's diagram (an AI write the tab missed, a cleared document
  that was saved) is shown and the tab's copy goes to History.
- Late answers to an old state's push or poll are dropped; a failed push
  says the server is unreachable; Download as .drawio saves the canvas.

Settings and server:
- Saved providers this version does not know stay in storage with their
  keys, and sending no longer trips over them.
- The desktop "Ollama (Local)" preset with a key goes to local Ollama
  again; a server model's Ollama URL variable is read; the admin panel
  writes Ollama Cloud's URL for a key without one.
- Provider error texts show again in the desktop app and for EdgeOne.
- .env: a quoted value followed by a comment ending in a quote is read as
  dotenv reads it; unquoted values are unchanged.
- Desktop app: the next launch opens the port where a chat was last
  saved; a launch elsewhere that saves nothing does not move it, and a
  page with no chats lets the next launch try the other port once.
- The Test button no longer stays busy after another tab changed the key.
- A completed append_diagram is no longer undone by an earlier failed
  edit's preview; a file read once in vain is saved again once it is read
  or gone.

From the first batch's review:
- The admin panel's Test of an entry without a URL now tests the server's
  <P>_BASE_URL, where chat sends the entry's key; chat is unchanged (the
  first fix rerouted working setups).
- The model list ends downloads that are too large, accepts answers
  without a body, and keeps the "redirects are not allowed" explanation.
- A test covers the preview's History rendering.

* fix: older defects (batch C) and the second batch's review

Chats:
- New Chat right after an answer saves that chat once. Saves run one at
  a time and read the chat on screen when their turn comes; a save
  scheduled for a chat that is no longer on screen is dropped. A chat
  whose id was still on its way to the URL no longer comes back after
  New Chat (the next answer went into it).
- Crossing the 768 px breakpoint keeps the chat panel: a streaming answer,
  unsaved messages and attachments stay. The panel gets the sizes of each
  side, and a panel collapsed on desktop opens on mobile.
- The chat's export waits for its own reply: an edit's history export
  still on its way no longer answers it with the older diagram, and two
  file saves at once no longer swap results.
- A second edit in one answer is previewed on the first edit's result.
- Stop also ends a running screenshot check; a chat that cannot be saved
  (storage full) can be left with "Continue without saving".
- Small diagrams with shapes count as diagrams; the tool card no longer
  crashes on malformed operations.

Quota and providers:
- Requests that reach the server's own endpoints count toward the quota:
  EdgeOne (always its own endpoint now), a private base URL whatever key
  header is sent, keyless Ollama without a URL. With the quota on, a
  redirect is followed only to a public address. The output cap applies
  to these requests too.
- Stop records the tokens of the steps that finished; the screenshot check
  counts its tokens without counting a request.
- EdgeOne configured only by AI_PROVIDER works, also in the admin Test,
  which forwards the access code. Azure set up only in the admin panel
  works in chat. The Test sends a Bedrock session token.
- The admin panel's Test of an entry without a URL uses the server's URL
  as the server does (no private address check for it); the admin panel
  no longer writes an Ollama URL.

MCP server:
- Write tools and start_session run one at a time, so two at once never
  drop each other's change; a cancelled call waiting its turn is skipped.
  get_diagram and export_diagram keep the session they started with.
- Export to .drawio first gets the user's latest edits from the browser.
- History thumbnails: one that arrives after the next AI write is
  dropped; a sync reply keeps the image; a version that changed only page
  settings is its own entry.
- A diagram over the 10 MB limit is saved without its image, or the user
  is told to download it (the server now answers 413 instead of cutting
  the connection).
- Labels holding text like id='1' or parent='1' are no longer read as
  attributes (a layer or a parent was deleted). A broken bare
  <mxGraphModel> file is refused.
- After a sync reply the tab no longer sends its autosave copy again.

Desktop and files:
- A newer switch of the same preset is not rolled back by an older one
  that failed. .env values with escaped quotes are read whole.
- MCP saved files: a file that could not be read stays protected while a
  folder without permission hides it, and is saved again once deleted.
- The desktop app reports "no chats" only when the count was read and no
  model settings are stored.

* fix: what the batch C review found

- A redirect followed for a custom base URL (quota on) no longer carries
  the user's key or cookies to another origin, as fetch itself does, and
  a private address may redirect to another private one (already counted).
- The admin Test of an Ollama or Vertex AI entry without a URL goes where
  chat sends that entry's key: the environment's own URL variable, for
  Ollama else the local default. The Test of an Ollama Cloud key without
  a URL went to the cloud while chat went to local Ollama.
- Chat saves: each save notes the chat on screen and the order of the
  reads before reading its data. A save read before switching chats no
  longer writes into the chat switched to, and a copy that waited for its
  thumbnail no longer replaces a newer one.
- "Continue without saving" keeps its button when a later auto-save fails,
  and goes away when a new message is sent.
- A screenshot check that was waiting for its image when the user pressed
  Stop stays skipped after the next message.
- MCP History: draw.io's own copy of a diagram (after get_diagram) no
  longer adds an entry without a picture; a change of background is still
  its own version. The tab ignores an edit's answer that arrives after a
  newer AI write loaded.
- Desktop: a deleted preset is not brought back by a failed switch, and a
  request naming no preset does not stop a rollback. An origin keeping
  an access code counts as having settings.
- .env: a quoted value ending in a backslash ("C:\dir\") is read as dotenv
  reads it.
- The tool card no longer crashes on an id that does not turn into text;
  an older Test's success timer no longer ends a newer Test's spinner.
- Tests that passed without their fix now check it.

* fix: what the whole-PR review and Copilot found

- A redirect followed for a custom base URL also drops the key headers of
  providers that do not use Authorization (x-api-key, x-goog-api-key,
  api-key) when it goes to another origin.
- A second Enter or click while a message is being prepared (attachments
  read, diagram exported) no longer sends it twice.
- The admin Test on the deployment's own endpoints (EdgeOne, the server's
  keyless Ollama, an address on the server's network) counts toward the
  quota like a chat; the chat and the Test share one rule for it. The
  Test of an Azure entry set up by AZURE_RESOURCE_NAME only goes where chat
  goes.
- EdgeOne's function is called at the site root again, as on main: EdgeOne
  serves edge functions there, outside Next's base path.
- MCP History: the state before a write is kept unless the browser saved
  no change of the user's since the last server write (draw.io's sync copy
  of it adds no entry), and the dedupe compares the exact text again, so a
  change of page size or other settings only is its own version.
- MCP: an edit keeps untouched labels as draw.io shows them (a literal line
  break in an attribute is a space); a new document of empty pages the
  user named is auto-saved; load_diagram reads only regular files, so a
  pipe cannot hold up the other write tools; the preview does not load
  back its own push still on its way (an undo made meanwhile is saved).
- Two overlapping saves of a new chat no longer reload the canvas from the
  older copy.
- At most three screenshot checks per user turn, passed or failed, as
  documented.
- Desktop: the main window navigates only within the app (draw.io stays in
  its frame); a presets file that is not JSON and cannot be moved aside is
  not overwritten.
- A last self-closing cell with a raw "<" in a value is not taken for cut
  off output.
- README: Material Design shapes load their icons from fonts.gstatic.com.

* fix(chat): stop saving and exporting an idle chat every second

Each auto-save takes a thumbnail, and draw.io's SVG gets a new random id
every time, so latestSvg changed and the diagram context rendered again.
getThumbnailSvg was a new function on every render, the auto-save depends
on it, so the next save was scheduled a second later: a chat with some
diagrams was saved and exported once a second for as long as it was open.
getThumbnailSvg (and the export helper it uses) now keep one identity;
they read refs only.

Found by hand in the production build; the new e2e test fails without the
fix on a production build (as CI runs it).

---------

Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
2026-10-05 22:39:32 +09:00

51 lines
1.9 KiB
TypeScript

/**
* Refuse a POST that a page on another website could have sent. A browser
* sends a cross-site POST without asking first (CORS preflight) only with a
* text or form body, so the routes take JSON only. In the desktop app also
* refuse a foreign Host: a site that points its own domain name at
* 127.0.0.1 (DNS rebinding) is same-origin with the local server, but its
* requests carry that domain. A request the server builds itself has no
* Host. Returns the response to send, or null when the request may go on.
*/
export function rejectCrossSite(req: Request): Response | null {
const contentType = req.headers.get("content-type") ?? ""
if (!/^\s*application\/json\b/i.test(contentType)) {
return Response.json(
{ error: "Content-Type must be application/json" },
{ status: 415 },
)
}
const host = req.headers.get("host")
if (
process.env.NEXT_AI_DRAWIO_DESKTOP === "1" &&
host &&
!/^(127\.0\.0\.1|localhost)(:\d+)?$/i.test(host)
) {
return Response.json({ error: "Forbidden" }, { status: 403 })
}
return null
}
/**
* Check the x-access-code header against ACCESS_CODE_LIST.
* Returns a 401 response to send back when the check fails, or null when the
* request may continue (including when no access codes are configured).
*/
export function checkAccessCode(req: Request): Response | null {
const accessCodes =
process.env.ACCESS_CODE_LIST?.split(",")
.map((code) => code.trim())
.filter(Boolean) || []
if (accessCodes.length === 0) return null
const accessCodeHeader = req.headers.get("x-access-code")
if (accessCodeHeader && accessCodes.includes(accessCodeHeader)) return null
return Response.json(
{
error: "Invalid or missing access code. Please configure it in Settings.",
},
{ status: 401 },
)
}