Better Auth guards `/unlink-account` with `freshSessionMiddleware`, which
rejects any session whose `createdAt` is older than `freshAge` (one day by
default). Sessions here last a week and there is no re-authentication flow to
refresh that timestamp, so disconnecting a provider failed with
`SESSION_NOT_FRESH` for every user who signed in more than a day ago.
Disable the freshness gate, and teach `getReadableErrorMessage` to read plain
error objects: Better Auth client errors are `{ code, message, status }`
objects rather than `Error` instances, so every auth toast was collapsing to
its generic fallback instead of showing the real reason.
* docs(adr): propose agent AI SDK v7 adoption plan
* fix(ai): bind analyzeResume through aiService in service test
The test destructured analyzeResume as a named export that does not exist; main was red.
* test(agent): keep pure ai helpers real via spread-actual mock factory
* feat(agent): add run guards, patch version guard, run wall-clock timeout
* feat(agent): validate UI messages at the send boundary
* feat(agent): crash-safe draft-row persistence and server-side cancellation
* feat(agent): reap stale run claims at boot, on send, and on thread open
* feat(agent): fresh-document patch output and tiered context pruning
* feat(ai): shared agent tool contracts and message metadata schema
* feat(agent): add per-thread review-patches setting with update endpoint
* feat(agent): gate resume patches behind hmac-signed tool approval
* feat(agent): merge question answers and approval decisions before run claim
* feat(agent): approval ui with composed auto-send and fixture-driven tests
* feat(agent): usage metadata, tool activity cards, smoother streaming
* feat(agent): tool-call repair, input examples, structured step logging
* chore(i18n): translate new agent workspace strings across all locales
* fix(agent): gate stale-run draft cancellation on winning the claim clear
Snapshot streaming drafts before the conditional clear and skip the flip entirely when another reaper or a replacement run already cleared the claim. Also address review nits in eleven locale catalogs.
* fix(agent): flip reaped drafts only when their snapshotted state is unchanged
* fix(agent): address review findings across run lifecycle, context budget, and approval flow
- bind patches to the revision the model read via signed baseUpdatedAt
- claim the run before consuming a continuation; recorded-but-unexecuted approvals retry as pending continuations
- keep run ownership on stop() until cancellation persists; preserve the claim for the reaper when final persistence fails
- estimate tokens without serializing binary attachments (tokenx) and enforce the budget by dropping oldest whole turns
- mark crash-recovered patch results as snapshot boundaries; strip /data prefixes at execution time
- retry failed continuations without regenerate; mount a single AgentChat; disable response controls on read-only threads; freeze review toggle during runs (client+server)
- accumulate usage across continuations and match the SDK's nested usage shape; label-form token strings; reorderable source label; accessible note field; state-neutral web-search label
* chore(i18n): translate revised agent strings across all locales
* fix(agent): harden baseUpdatedAt validation and address review follow-ups
- bundle tokenx in the server runtime dependencies (e2e boot failure)
- strict ISO schema for baseUpdatedAt plus loud executor rejection of unparseable values
- it-IT source label consistency (Fonte)
- prove penultimate-turn retention in the context pruning test
* chore(deps): exempt tokenx from knip for the externalized server bundle
Fills the 25 strings added this cycle by the toast migration, the account page
rename, the Custom Styles status labels and the job posting auto-fill, across
all 53 target catalogs. The zu-ZA pseudo-locale is intentionally left empty.
Adds the questionnaire and empty-state primitives and renders the
ask_user_question tool call inline in the chat, so the agent can offer choices
instead of guessing when a request is ambiguous.
Fetching an arbitrary job URL server side meant owning SSRF defence, redirect
and size limits, and per-site scraping quirks. The autofill tool now takes only
pasted text, so the URL input, the fetch path and its MCP annotation are gone.
The sheet gates the call behind a tested AI provider and a minimum paste length
so a stray snippet does not spend an AI call.
The page now holds account-level actions rather than only destructive ones, so
it is reachable at /dashboard/settings/account and presented with a neutral
icon in the sidebar and command palette.
Swaps sonner's toast.success/error/loading/dismiss for the new toast.add({ type,
description }) and toast.close across dialogs, auth pages, the builder, the
dashboard and the applications views. Behaviour is unchanged.
Each case rebuilt the whole spec, which walks every router and resume JSON
schema. The first case already carried a raised 15s timeout and still timed out
on a loaded machine. The spec is deterministic and only read here, so build it
once: the file drops from over 15s to 1.86s.
The case passed cursor position 5 into "--resume-", which lands mid-token and
reads as a selector context, so it received the selector list. Every other case
in the file uses source.length.
Disabling the textarea for the duration of a response made the browser blur it,
so the caret left the composer on every send and had to be clicked back. send()
already ignores calls mid-stream, so Enter stays a no-op and type-ahead works.
Every step-start part serialises to the same JSON, so the content-derived key
collided for any multi-step assistant message and React warned about duplicate
keys on each incoming chunk. Two identical text parts collided the same way.
Parts are append-only and never reordered by the AI SDK, so the index is stable.
Public resume pages only produced their OpenGraph and Twitter tags client side,
so a shared link had no card at all. The server now injects them into the shell
and swaps in the resume's own title and description.
The lookup is scoped to public, password-free resumes and deliberately avoids
resumeService.getBySlug: that counts a view and would expose a protected
resume's summary to an unauthenticated crawler. User-authored values are escaped
before they reach the HTML, and any lookup failure falls back to the plain shell.
getResumeSocialMeta is shared with the client route head so the two cannot drift.
X reads twitter:* meta tags from the name attribute, not property, so the card
validator reported twitter:title and twitter:description as missing. Also adds
the og:type tag the root head was never emitting.
The static middleware was mounted ahead of the web app fallback, and Hono's
serveStatic resolves "/" to the directory and returns dist/index.html verbatim.
handleWebApp never ran for the root route, so the OpenGraph, Twitter, canonical
and JSON-LD markup it injects was missing in production - fetching
https://rxresu.me/ as Twitterbot returned zero og: tags.
Route "/" explicitly before the static middleware so the injection runs.
* feat(resume): add a deterministic ATS parseability check
Adds an offline ATS linter that reports whether a parser can read a
resume, surfaced as an always-on panel in the builder.
The existing Resume Analysis panel needs a configured AI provider, so
users who never set one up get no feedback at all. These 22 rules run
as a pure function over ResumeData with no provider, no network and no
rendered PDF, so they work for everyone on every edit.
Rules cover contact details, date parseability, sections that hold
content but never render, column and sidebar placement, and typography
thresholds. The catalog mirrors the Semantic CSS diagnostic catalog:
stable codes carrying a severity, meaning and action, with no i18n
dependency so the web layer translates by code. Each finding carries a
JSON Pointer, which is what makes jump-to-field work.
Deliberately no second score. Resume Analysis owns overallScore, so
this reports "N of M checks passed" and counts by severity instead.
* fix(resume): accept localized ongoing periods and reject bare ones
Two period-parsing bugs found in review.
The ongoing-token set was English-only, so a German resume reading
"2020 - heute" was reported as unparseable and the panel told the user
to rewrite a perfectly valid range. Rather than guess translations for
55 locales, a range ending that carries no digits and is not a month
name in the resume's locale is now read as ongoing. That keeps a
genuinely incomplete ending such as "Jan 2020 - Feb" reported, since
"Feb" resolves as a month.
A bare "Present" also parsed as a valid period, so an experience entry
with no start date passed the check. A standalone ongoing token is now
rejected; ongoing tokens remain valid as the end of a range.
* feat(web): scroll ATS findings to the item they belong to
Findings for different items in one section all landed on the section
header, so a date problem on the third role gave no more help than
naming the section.
getAtsFindingTarget now resolves the offending item from the finding's
JSON Pointer against the resume, and SectionItem carries a matching DOM
id. The panel scrolls to that item and falls back to the section header
when the item is not mounted, which is what happens while its section
is collapsed.
* fix(resume): recognize ongoing periods by token, not by shape
The previous heuristic read any short, digit-free range ending as an
ongoing marker, so "2020 - unknown", "2020 - later" and "2020 - tbd"
parsed cleanly and suppressed the finding they should have raised.
Replaced with an explicit table of ongoing words keyed by language,
covering the locales the app ships. Matching is exact, so unrecognized
endings are reported again. A locale missing from the table falls back
to the earlier behaviour of reporting its ongoing periods, which is a
visible gap someone can close by adding a word rather than a silent
hole in detection.
Tests assert every listed token parses and that the table stays
lowercase, since lookups normalize that way.
* fix(ats): parse punctuated ongoing tokens
* fix(ats): parse Unicode punctuated ongoing tokens
---------
Co-authored-by: Amruth Pillai <im.amruth@gmail.com>
* fix(import): auto-detect JSON format and show readable errors (#1)
Readable import errors, a fail-soft v4 parser, and auto-detect of the JSON format so uploads just work. The format dropdown becomes an optional override. PDF and DOCX (AI) paths are untouched.
* fix(import): address review feedback on the v4 guard and error message
- reactive-resume-v4-json.tsx: reject arrays in isRecord so array-valued
basics, sections, or metadata no longer pass the v4 shape guard.
- reactive-resume-v4-json.tsx: reuse the guard's error instance in the
catch arm instead of allocating a duplicate NOT_V4_MESSAGE.
- error.ts: use a singular "Problem" label for root-level Zod issues so
the message stays grammatical.
- add a regression test for array-valued v4 branches.
* fix(import): preserve selected JSON format
* test(import): cover selected JSON parser
---------
Co-authored-by: MrTig-afk <MrTig-afk@users.noreply.github.com>
Co-authored-by: Amruth Pillai <im.amruth@gmail.com>
react-zoom-pan-pinch calls preventDefault() on its window-level mousedown
listener so that dragging the canvas does not select text. That also cancels the
browser's native focus shift, so focus stays wherever it was before the pan --
typically the sidebar button that opened the last dialog, since closing a dialog
restores focus to its trigger. A focused button activates on Space keyup, so
holding space to pan and then releasing it reopened the most recent dialog.
Blur the focused element from onPanningStart, which reinstates exactly the focus
change the browser would have made on its own. onPanningStart only fires when the
mousedown target is inside the transform wrapper, so sidebar and dialog clicks are
unaffected, and keyboard-only users never trigger a pointer pan.
Closes#3300
Co-authored-by: Amruth Pillai <im.amruth@gmail.com>
The icon picker grid starts with an empty string entry that renders the
"no icon" (prohibit) symbol, but the onClick guard "if (icon)" treated
the empty string as falsy and ignored the click. Change the guard to
check for a defined string value so the empty/no-icon option can be
selected.
Closes#3252Closes#3261
Co-authored-by: Devin <devin@example.com>
Co-authored-by: Amruth Pillai <im.amruth@gmail.com>
* fix(auth): use loopback URL for MCP OAuth JWKS verification
Fetch the JWKS endpoint over the internal loopback address instead of the public APP_URL, so token verification works under Docker port-mapping, reverse proxies, and other deployments where the public URL does not loop back to the Node process.
Also log the specific MCP OAuth verification error instead of swallowing it with a bare catch.
Fixes#3077
* fix(auth): normalize internal JWKS URL and throttle MCP OAuth warnings
- Problem: default loopback JWKS URL used PORT in dev where the server
listens on SERVER_PORT (3001), and trailing-slash overrides produced
//api/auth/jwks; unthrottled warn logs could flood on bad bearer tokens.
- Fix: resolveInternalBaseUrl trims/normalizes BETTER_AUTH_INTERNAL_URL,
mirrors apps/server listen-port selection, and MCP OAuth warnings are
throttled to once per minute.
- Verification: pnpm exec biome check on changed files; pnpm typecheck.
* fix(auth): declare BETTER_AUTH_INTERNAL_URL in turbo globalEnv
- Problem: Turborepo strict env mode strips undeclared BETTER_AUTH_INTERNAL_URL under pnpm dev, so the JWKS override silently falls back to loopback.
- Fix: add BETTER_AUTH_INTERNAL_URL to turbo.json globalEnv (required for any new env var per CLAUDE.md).
- Verification: python3 JSON parse of turbo.json; confirmed var was absent from globalEnv before this change.
---------
Co-authored-by: Amruth Pillai <im.amruth@gmail.com>
Node 24 terminates the whole process on an unhandled promise rejection, so a
single request's stray rejection could take the server down for every user
(as the USER_STOPPED agent-abort bug did). Add a process-level unhandledRejection
handler that logs and keeps serving. Uncaught exceptions are intentionally left
on Node's default crash-and-restart, since process state is unsafe afterward.
Claude-Session: https://claude.ai/code/session_01ULhhLQ24DvnYwzP4afDuye
- Add an sr-only "Copy secret" label to the icon-only copy button in the 2FA
enable dialog; it was previously announced as an unlabeled button.
- Add a focus-visible ring to template gallery cards. The only ring was gated
on the selected state, so keyboard focus was invisible while tabbing.
- Disable the API-key create submit button while the request is in flight to
prevent duplicate keys from a double-click.
Surfaced by a shadscan UI audit. The remaining ~95 findings were false
positives from the auditor not understanding the pnpm monorepo and the
TanStack Start root-route shell, and were waived.
Claude-Session: https://claude.ai/code/session_01JYTniVDeA56o1kGhdoCUoD
The server PDF preflight spawned a fresh worker per semantic-CSS edit, each
racing a 15s startup deadline to cold-load the ~721kB+5MB PDF runtime. That
load is super-linear in CPU (~3s at 1 vCPU, >15s on a throttled/shared vCPU),
so on a constrained box every edit hit the startup-timeout path and returned
STYLESHEET_PREFLIGHT_WORKER_FAILED. Since the service only advances the applied
stylesheet when preflight passes, custom styles never applied and the editor
stuck on Checking.
Warm one worker at boot and reuse it (message-based input, respawn on
crash/timeout), so the cold load is paid once instead of per edit. Raise the
render deadline 5s->30s (a rich resume renders ~5-18s on a slow box) and the
readiness ceiling to 120s so the one-time warm completes even when throttled.
Surface worker load failures instead of an unhandled-rejection crash, and log
runner-side failure paths so the previously opaque failure is diagnosable.
Verified in node:24-slim under --cpus=0.25/0.35/0.5: all reused requests pass.
Claude-Session: https://claude.ai/code/session_01ULhhLQ24DvnYwzP4afDuye
Concurrent compile requests from editor intelligence were rejecting
in-flight edit compiles as stale, and the store swallowed that rejection
without leaving compiling. Resolve all compile results and surface
compile failures as an error status so styles can apply again.
Co-authored-by: Cursor Agent <cursoragent@cursor.com>
github-actions[bot]andCrowdin Bot
Amruth Pillai
ignaciocarreandAmruth Pillai
Syed Ali Abbas ZaidiandAmruth Pillai
Santhi PrakashandAmruth Pillai
autofix-ci[bot]
