* docs: document model/auth settings, /model --vision, and --safe-mode
Refresh user docs to match the current codebase:
- commands.md: add the /model --vision override (vision-bridge model)
- settings.md: add model.baseUrl, model.sessionTokenLimit, visionModel,
and voiceModel; document the deprecated security.auth.apiKey and
security.auth.baseUrl keys with a pointer to modelProviders
- troubleshooting.md: document the --safe-mode flag for isolating
customization issues
* docs: address review feedback on sessionTokenLimit, safe-mode, deprecation notes
- model.sessionTokenLimit: correct default to -1 (runtime fallback in
core/config.ts) and clarify breach behavior (current send dropped, not
session abort) per client.ts SessionTokenLimitExceeded handling.
- --safe-mode: expand the disabled-customizations list to also cover
permission rules, approval mode overrides, memory features, and sandbox
settings, matching cli/config.ts.
- security.auth.apiKey/baseUrl: align deprecation wording with the existing
tools.* entries (**Deprecated.**) and drop the unsubstantiated
'(slated for removal)' qualifier.
* docs: note QWEN_CODE_SAFE_MODE env var as a safe-mode alternative
Document the QWEN_CODE_SAFE_MODE=true environment variable as an
alternative activation path for safe mode, for cases where the CLI
cannot accept flags (verified against isSafeModeEnv in
packages/core/src/utils/safe-mode.ts).
* docs: clarify model.baseUrl, sessionTokenLimit=0, and safe-mode subagents
- model.baseUrl: describe it as a picker-managed disambiguator, not a
hand-editable override (stale values can misroute to a same-id provider).
- model.sessionTokenLimit: note that 0 is treated as unlimited (same as -1),
unlike model.maxToolCalls where 0 disallows all calls.
- --safe-mode: include custom subagents in the list of disabled
customizations (only built-in subagents load in safe mode).
* docs: clarify sessionTokenLimit semantics and add --safe-mode to headless flags
- settings.md: reword model.sessionTokenLimit to reflect that the gate
compares the last recorded prompt token count before the next send
(not a per-send preflight cap), and that the next send is dropped.
- headless.md: add a --safe-mode row to the CLI flags table so the
diagnostic flag is discoverable there, cross-referencing Troubleshooting.
* docs: align safe-mode sandbox wording to 'sandbox settings'
Safe mode passes an empty Settings object to loadSandboxConfig
(packages/cli/src/config/config.ts:1793), so it strips settings-sourced
sandbox config while the --sandbox flag and QWEN_SANDBOX env still apply.
Match headless.md to troubleshooting.md's accurate 'sandbox settings'.
* docs: correct safe-mode approval-mode wording and align both lists
Safe mode only strips settings-sourced approval mode; the --yolo and
--approval-mode CLI flags are evaluated before the safeMode guard
(packages/cli/src/config/config.ts:1521-1528) and still take effect.
Reword to 'settings-sourced approval mode overrides' and note the CLI
flags in troubleshooting.md and headless.md, and make the enumerated
safe-mode disable list identical (same items and order) across both.
---------
Co-authored-by: Claude <noreply@anthropic.com>
* feat(core): add --insecure flag to skip TLS verification for self-signed endpoints (#3535)
Enable skipping TLS certificate verification for outbound model API connections via a new --insecure CLI flag, QWEN_TLS_INSECURE, or NODE_TLS_REJECT_UNAUTHORIZED=0. The setting is applied to the undici dispatcher Qwen Code installs: a direct connection uses connect TLS options, while a proxied connection disables verification for the upstream origin (requestTls) and a self-signed HTTPS proxy (proxyTls). Off by default; behavior is unchanged when not enabled.
Fixes#3535
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* fix(core): address review on --insecure (security + simplification) (#3535)
- Block project .env from enabling QWEN_TLS_INSECURE by adding it to PROJECT_ENV_HARDCODED_EXCLUSIONS, so an untrusted repo cannot silently disable TLS verification for all API connections.
- Remove the unverifiable Bun fetch `tls` special-casing; Bun users can still opt out via NODE_TLS_REJECT_UNAUTHORIZED=0, which Bun honors natively.
- Do not suggest `--insecure` in the TLS error hint when verification is already disabled; show a network/protocol-oriented message instead.
- Add tests: env-flag regex/falsy branches, loadCliConfig env side-effect, fetch hint variants, and a security guard for the project .env exclusion.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* fix(core): harden --insecure per review (#3535)
- Block NODE_TLS_REJECT_UNAUTHORIZED from project .env too (initial load only consults PROJECT_ENV_HARDCODED_EXCLUSIONS), since isTlsVerificationDisabled() honors it.
- When opting out, set NODE_TLS_REJECT_UNAUTHORIZED=0 process-wide in loadCliConfig and emit a stderr MITM warning. This makes the opt-out effective on the Bun runtime and the proxy-creation fallback path (which the undici dispatcher does not cover), and gives a user-visible signal.
- Avoid evaluating isTlsVerificationDisabled() twice on the proxy path via a default parameter on getOrCreateSharedDispatcher.
- Update tests accordingly.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* fix(cli): broaden --insecure warning and log it; add env-path tests (#3535)
- Broaden the TLS-disabled warning to state the process-wide blast radius (API, OAuth, MCP servers, child processes), since NODE_TLS_REJECT_UNAUTHORIZED=0 is set process-wide.
- Also emit the warning via debugLogger so the state is discoverable in ~/.qwen/debug/ after terminal scrollback is gone.
- Add tests for the env-var-only path (pre-set QWEN_TLS_INSECURE) and the already-disabled guard (no duplicate assignment/warning).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
* docs: add CLI subcommands section with qwen sessions list
- Add section 5 to commands.md for CLI-level subcommands
- Document qwen sessions list with --json and --limit flags
- Include output format, examples, and usage patterns
The sessions list command was added in commit 14e6ae8c2 but not documented.
* docs: add vertex-ai auth, missing commands, and qc-helper index entries
Audit docs/ against current codebase and fix high-impact drift:
- Add vertex-ai to auth.md, model-providers.md, and tos-privacy.md supported
auth type tables (was missing from all three)
- Add missing slash commands to commands.md: /cd, /import-config, /workflows
- Add /doctor subcommands (memory, cpu-profile, rollback) and /extensions
subcommands (list, manage, explore, install) to commands.md
- Add undocumented altNames: /clear→/reset,/new; /stats→/usage;
/auth→/connect,/login; /resume→/continue; /compress→/summarize
- Add 8 missing feature entries to qc-helper SKILL.md topic index:
code-review, followup-suggestions, tool-use-summaries, markdown-rendering,
structured-output, dual-output, channels, tips
- Fix CLI binary name in contributing.md (qwen-code → qwen)
* docs: resolve review feedback on auth count, /workflows usage, GOOGLE_MODEL example
- auth.md: correct intro count from 'four' to 'three' methods (3 bullets + 3 Options; Vertex AI is a provider under API Key)
- commands.md: add '/workflows <runId>' to usage column to match argumentHint '[runId]'
- model-providers.md: add required GOOGLE_MODEL to the Vertex AI env example so it matches the prose and modelConfigUtils requirement
* docs: remove duplicate /workflows row introduced by main merge
main already documents /workflows (with the <runId> usage); the branch
merge kept both rows, leaving a duplicate in the Tool and Model
Management table. Drop the redundant row added by this PR.
* docs: resolve review feedback — /extensions explore source arg, CLI-name remnants
- commands.md: /extensions explore requires a <source> (exploreAction errors 'Unknown extensions source' without it)
- contributing.md: finish the qwen-code -> qwen CLI rename on the debug note (binary + .qwen config dir); repo-name references left intact
* docs: resolve review feedback — Vertex AI in tos-privacy, /doctor rollback clarity
- tos-privacy.md: propagate Vertex AI (which the header already counts as the 4th method) into the Data Collection list, FAQ Q1 and Q3, and add a '4. If you are using Vertex AI' section pointing to Google Cloud terms — fixes the four-vs-three internal inconsistency
- commands.md: clarify /doctor rollback rolls back the standalone CLI binary (standalone installs only) and disambiguate from /rewind's rollback alias (doctorCommand.ts gates on isStandalone)
* docs: resolve review feedback — clear semantics, doctor argHints, arrow spacing
- /clear: fix description to 'Clear conversation history and free up context' and drop the misleading '(shortcut: Ctrl+L)' grouping; Ctrl+L only clears the screen (clearScreen), it does not reset the session like /clear (clearCommand.ts)
- Ctrl/cmd+L keyboard row: clarify it clears the visible screen only, not 'Equivalent to /clear'
- /doctor memory and /doctor cpu-profile: surface the full argumentHints ([--sample] [--snapshot], [--duration <seconds>]) from doctorCommand.ts
- normalize section 1.4 arrow subcommands to spaced '→ ' style (approval-mode rows were the lone outliers)
* docs: resolve review feedback — enumerate /extensions explore sources
List the two valid sources (Gemini, ClaudeCode) from EXTENSION_EXPLORE_URL in extensionsCommand.ts so users can discover them without trial and error.
* docs: resolve review feedback — add extensions install security warning
/extensions install (extensionsCommand.ts) installs arbitrary git repos/paths with no confirmation prompt; add a warning that extensions run with full Qwen Code permissions and should only come from trusted sources.
* docs: resolve review feedback — add /stats subcommands, /auth aliases in 1.11
- commands.md: add /stats daily, /stats monthly, /stats export rows (registered in statsCommand.ts with day/month aliases and --format csv|json)
- section 1.11: note /auth's /connect and /login aliases (parallel to section 1.4)
* docs: resolve review feedback — /stats export full args, /summarize note
- /stats export: show the full argumentHint ([date|month] and [--output path]) from statsCommand.ts
- add a note disambiguating /summarize (alias of /compress, destructive) from /summary (project summary)
* docs: resolve review feedback — complete /arena /ide /directory /voice /mcp usage
Add the missing subcommands/arguments shown in the command sources:
- /arena: stop, select (arenaCommand.ts)
- /ide: enable, disable (ideCommand.ts)
- /directory: show (directoryCommand.tsx)
- /voice: hold, tap, off (voice-command.ts argumentHint)
- /mcp: nodesc, schema, auth, noauth (mcpCommand.ts argumentHint)
* docs: resolve review feedback — arena/stats aliases, trim /stats description
- /arena select: note alias 'choose' (arenaCommand.ts)
- /stats daily, /stats monthly: label day/month as aliases (statsCommand.ts)
- /stats: trim the description to a terse behavior-focused line (drop volatile tab names/keyboard shortcuts that belong in the dashboard help)
* docs: resolve review feedback — /copy args, /doctor memory --snapshot warning
- /copy: document language/latex/mermaid/index selection (copyCommand.ts argumentHint)
- add a warning that /doctor memory --snapshot writes a heap snapshot with sensitive data (matches doctorCommand.ts runtime warning)
* docs: resolve review feedback — mcp/approval-mode/copy accuracy, qwen privacy URL
- /mcp: drop deprecated auth/noauth (mcpCommand.ts argumentHint is now desc|nodesc|schema; auth/noauth are stubs)
- /approval-mode: drop nonexistent --project (mode is session-only), show actual invocations, add a safety warning for auto-edit/auto/yolo
- /copy: note N = Nth-last reply (copyCommand.ts)
- tos-privacy: unify Qwen Privacy Policy URL to qwen.ai/privacypolicy
* docs: resolve review feedback — import-config args + feature-gated commands note
- /import-config: show 'all' (default source) and enumerate --scope user|project (importConfigCommand.ts)
- add a note that /workflows, /lsp, /trust register only when their feature setting is enabled (BuiltinCommandLoader.ts gates them, default off)
* docs: fix feature-gating mechanisms + restore /stats tab names
- Correct the /workflows/lsp/trust note: actual gates are QWEN_CODE_ENABLE_WORKFLOWS=1 (env),
--experimental-lsp (CLI flag), and security.folderTrust.enabled (setting) — the prior
workflowsEnabled/lsp.enabled/folderTrust keys did not exist
- /stats: restore the Session/Activity/Efficiency tab names (dashboard contents are not
documented elsewhere); keep volatile keyboard hints out per the earlier review
---------
Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com>
* fix(installer): tighten verifier base-url + clarify test helper
Three small refinements from the second review pass:
- normalizeHttpsBaseUrl rejects everything except https, since real release
URLs are always HTTPS. Accepting http previously would let an operator
silently target a stale or attacker-controlled mirror.
- Drop EXPECTED_RELEASE_ASSET_NAMES from the public exports; it was only
used internally for the verification log line.
- Rename the test helper standaloneChecksumContent to
placeholderChecksumContent and document that the hashes in its output are
placeholders — the remote verifier does not download archives or compare
hashes, it only validates that SHA256SUMS lists the expected names and
that each archive URL is reachable.
The non-https rejection test now also covers `http://` in addition to the
existing `file://` case.
* style(installer): align installer completion output
* revert(installer): keep hosted installer output unchanged
* fix(installer): address release validation review feedback
* docs: switch public install commands to standalone hosted entrypoint
Update README, quickstart, and overview to point at the new
install-qwen-standalone.sh / install-qwen-standalone.ps1 hosted URLs.
Add standalone uninstall instructions to Uninstall.md. Remove the
staged-rollout note from INSTALLATION_GUIDE.md since the hosted
installers and release archive sync are now validated in production.
* docs: clarify pull request size guidance
* fix(installation): harden standalone release validation
* fix(installation): redact release verifier credentials
* feat(installer): add visual branding to Linux/macOS install script
Add brand-colored ASCII art logo, custom download progress bar with
Unicode block characters, and step indicators [1/3] [2/3] [3/3] to
match the quality of competing CLI installers.
* fix(test): update stale assertion after guide text was removed
The text "Public installation documentation" was removed in 20f5243f6
but the test assertion was not updated, causing a persistent failure.
* feat(installer): use truecolor per-character gradient for logo branding
Replace 256-color block coloring with 24-bit truecolor per-character
gradient interpolation matching the CLI's ink-gradient rendering.
Colors follow the fallback gradient: #4796E4 → #847ACE → #C3677F.
Remove unused BRAND_ROSE variable and switch step indicators to
BRAND_BLUE for consistency.
* fix(installer): address critical review findings on SSRF, semver, and reliability
- Fix IPv4-mapped IPv6 SSRF bypass: handle 3-part hex representations
that Node.js produces (e.g. ::ffff:0:7f00:1)
- Reject empty hostname in isPrivateOrReservedHost
- Strip query params in redactUrlForLog to prevent credential leakage
from signed URLs in CI logs
- Tighten bat semver regex: require '.' or '-' separator before suffix
(rejects 1.2.3foo, matches shell installer behavior)
- Add -f flag to curl in download_with_progress so HTTP errors aren't
silently written as file content
- Restore terminal cursor in INT/TERM signal handlers (RETURN trap
doesn't fire on exit)
- Add unit tests for isPrivateOrReservedHost and redactUrlForLog
- Update test assertion for new split-pattern semver validation
* fix(installer): close release validation review gaps
* test(installer): cover shadowed qwen installs
* fix(installer): avoid npm auto-update for standalone installs
* fix(installer): block IPv4-compatible IPv6 SSRF and harden archive validation [skip ci]
- Add ipv4FromCompatibleIpv6() to detect deprecated RFC 4291 §2.5.5.1
addresses (e.g. ::7f00:1 → 127.0.0.1) that bypass SSRF protection
- Extend archive validation to reject hardlinks in addition to symlinks
- Add signal trap suppression during critical mv swap to prevent
partial-install state on Ctrl+C
- Add diagnostic logging to silent catch in standalone detection
* fix(installer): finish standalone install follow-ups
* feat(installer): streamline output with custom progress bar and minimal UX
Replicate OpenCode-style installer experience:
- Add custom ■-character progress bar with percentage (file-size polling)
- Remove verbose INFO:/SUCCESS: prefixes on happy path
- Simplify --help output to essential options
- Keep gradient logo, shadowing warnings, and PATH conflict detection
- Silence mirror probing, checksum, and npm detection messages
- Add "For more information" link to final output
Both .sh and .bat scripts updated consistently.
All 95 tests pass.
* feat(installer): add progress bar and logo to Windows installer
- Add PrintLogo subroutine with QWEN CODE ASCII header
- Add PrintProgressComplete using PowerShell VT100 ■-bar at 100%
- Show progress complete after successful download
- Add spacing in PrintHeader for consistent look with .sh
* fix(installer): address review findings on progress bar
- Replace `sleep 0.3` with `sleep 1` for busybox/minimal env compatibility
- Add file_size > 0 guard to avoid progress bar flicker on empty file
- Remove trailing blank lines before closing braces in 4 functions
* fix(installer): finalize Windows UX — suppress curl progress, fix logo
- Windows: suppress curl ### progress with -s --show-error (keep -#fSLo for test compat)
- Windows: use simple colored "Q W E N C O D E" logo (truecolor VT100)
- Windows: SHA256SUMS download uses DownloadFileQuiet (no progress bar for small files)
- Windows: remove SUCCESS/INFO PATH messages from MaybeUpdateUserPath
- Linux: fix double 100% progress bar (skip bar for files < 100KB)
* fix(installer): handle Windows backslash paths in standalone detection
`fs.realpathSync` returns backslash paths on Windows (e.g. C:\Users\...\lib\cli.js).
Normalize to forward slashes before matching the /lib/cli.js suffix so standalone
install detection works correctly on Windows.
Fixes CI: Test (windows-latest, Node 22.x)
* fix(installer): normalize expected paths in Windows standalone test
The existsSync mock built expected paths with path.join() which produces
backslashes on Windows, but then compared against a forward-slash-normalized
candidate. Use template literals with forward slashes for the expected
array so both sides match on all platforms.
* refactor(installer): simplify post-install output
Remove verbose post-install messages (install path, uninstall command,
PATH conflict warnings, npm coexistence tips) and replace with a clean
4-line summary matching OpenCode's minimal style.
* refactor(installer): simplify Windows post-install output
Match the Linux/macOS installer simplification — remove verbose
messages (install path, uninstall command, PATH warnings) and keep
only the essential 4-line success summary.
* refactor(installer): suppress verbose Windows messages
Remove "User PATH already starts with", backup WARNING messages,
and PS1 wrapper "Run: qwen" / "qwen is ready to use" output to
match the minimal Linux installer style.
* fix(test): align install-script assertions with simplified output format
The installer scripts were refactored to use a compact output format
(no separate To start/Installed to/Uninstall lines, no shadow warnings),
but the test assertions were not updated accordingly.
* fix(installer): align hardlink detection and expand test coverage
- Rename archive_contains_symlinks to archive_contains_symlinks_or_hardlinks
in install-qwen-with-source.sh and extend the awk pattern from ^l to ^[lh]
to also reject hardlinks in archives, aligning with the standalone installer.
- Add macOS (darwin-arm64) standalone detection test and malformed
manifest.json fallback test in installationInfo.test.ts.
- Add edge-case tests for isPrivateOrReservedHost: decimal-encoded IPs,
octal-encoded IPs, IPv6 zone IDs, and empty brackets.
* docs: update authentication methods to reflect OAuth discontinuation
Remove deprecated Qwen OAuth references and update documentation to
direct users to valid authentication methods (API Key, Coding Plan,
or Local Inference) following the OAuth free tier discontinuation on
2026-04-15.
Closes#3316
* docs: fix quickstart auth description to match actual /auth UI
The /auth command shows three options: Alibaba Cloud Coding Plan,
API Key, and Qwen OAuth (discontinued). Updated quickstart.md to
accurately reflect this UI instead of splitting into Option A/B/C.
Also updated settings.md, commands.md, and troubleshooting.md with
minor OAuth-related cleanups.
* docs: update .qwen workspace description in quickstart
Remove reference to 'Qwen account' since OAuth is discontinued.
The .qwen directory is created by Qwen Code itself for storing
credentials, configuration, and session data.
* docs: fix warning block formatting in quickstart
- Add missing '>' continuation for the OAuth discontinuation warning block
Co-authored-by: Qwen-Coder <qwen-coder@alibabacloud.com>
* docs: update README Qwen3.6-Plus description
- Remove mention of running Qwen3.6-Plus locally via Ollama/vLLM
- Keep only the Alibaba Cloud ModelStudio API key option
Co-authored-by: Qwen-Coder <qwen-coder@alibabacloud.com>
* docs: address review feedback - remove Local Inference from auth, add dual-region links
- Local Inference removed from auth method lists, kept as separate
'Local Model Setup' section with detailed Ollama/vLLM config examples
- All links now provide dual-region URLs (Beijing + intl)
- .qwen workspace note restored to original meaning (cost tracking)
- Device auth flow error kept scoped to legacy OAuth
- API setup guide links updated with confirmed intl URL
---------
Co-authored-by: Qwen-Coder <qwen-coder@alibabacloud.com>
* feat(auth): discontinue Qwen OAuth free tier (2026-04-15 cutoff)
The Qwen OAuth free tier has reached its end-of-life date. This updates
all client-side messaging, blocks new OAuth signups, and guides existing
users to alternative providers.
* fix(test): add getModelsConfig mock and update QWEN_OAUTH test expectations
- Add getModelsConfig() to Config mocks in gemini.test.tsx (3 failures)
- Update validateNonInterActiveAuth test to expect exit for QWEN_OAUTH
since validateAuthMethod now returns an error for discontinued free tier
* docs: update quota exceeded alternatives to OpenRouter and Fireworks
- Update README.md news section to recommend OpenRouter and Fireworks
as primary alternatives, with ModelStudio as third option
- Update retry.ts quota error message to include OpenRouter and
Fireworks URLs for users whose OAuth quota has been exhausted
Co-authored-by: Qwen-Coder <qwen-coder@alibabacloud.com>
* fix(test): update retry test assertions to match new quota error message
* docs: update free tier quota to 100 req/day with sunset notice and alternatives
Update all references to reflect the Qwen OAuth free tier policy change:
- 1,000 → 100 requests/day across code, i18n, and docs
- Add 2026-04-15 sunset date everywhere
- Guide users to OpenRouter, Fireworks AI, or ModelStudio in docs
- Remove CHANGELOG.md
---------
Co-authored-by: Qwen-Coder <qwen-coder@alibabacloud.com>
Co-authored-by: tanzhenxin <tanzhenxing1987@gmail.com>
- Replace nested API-KEY submenu with flat three-option layout
- Add descriptive labels for each authentication method:
- Qwen OAuth: Free, up to 1,000 requests/day
- Alibaba Cloud Coding Plan: Paid, multiple model providers
- API Key: Bring your own API key
- Simplify region selection for Coding Plan (China vs Global)
- Use DescriptiveRadioButtonSelect for better visual hierarchy
- Add itemGap prop to BaseSelectionList for spacing
- Update i18n strings in en.js, zh.js, and ru.js
- Simplify custom API key configuration info view
- Clean up unused region-specific strings
Closes#2016
Co-authored-by: Qwen-Coder <qwen-coder@alibabacloud.com>
