Quota fetcher:
- loadCodeAssist headers now match antigravity.go exactly
- fetchAvailableModels uses empty body {} and correct User-Agent
- Prevents accounts being flagged for anomalous requests
Error logs:
- Extract status code from end of log (RESPONSE section)
- Display full model names instead of abbreviated
- Show status badges (500/429/4xx) with color coding
- Add provider icons with white background
- Refresh token 5 minutes before expiry (matches CLIProxyAPIPlus)
- Only refresh when: expired, no expiry info, or expiring within 5 min
- Reduces unnecessary OAuth API calls when token is still valid
- Add isUnprovisioned flag to QuotaResult interface
- Detect when account is authenticated but lacks project ID
- Show actionable message: "Sign in to Antigravity app to activate quota."
- Refactor getProjectId retry logic for cleaner error propagation
CCS can now fetch quota independently of CLIProxyAPI by refreshing
access tokens using the refresh_token stored in auth files.
- Add refreshAccessToken() using Google OAuth token endpoint
- Use public Antigravity OAuth credentials (from CLIProxyAPIPlus)
- Proactively refresh token before API calls to avoid stale tokens
- Fallback retry on auth errors
This fixes quota showing "Access token expired" for accounts whose
file-based access_token is stale (CLIProxyAPI refreshes at runtime
but doesn't persist to disk).
Backend was returning 'Token expired' error based on stale file state
without attempting API call. CLIProxyAPIPlus refreshes tokens at runtime
but intentionally doesn't persist to disk, making file-based expiration
checks always misleading.
Now quota fetcher attempts API call regardless of file expiration state.
If token is truly invalid, API returns 401 which is handled properly.
- Remove "Token expired" warning (showed stale file state, not runtime)
- Add "Active/Last used" status based on CLIProxyAPI runtime stats
- Show green checkmark for recently used accounts (within 1h)
- Show "Not used yet" for accounts without usage stats
- Remove expired warning from flow-viz account cards
- Add model quota sorting (Claude > Gemini > GPT > other)
- Add quota reset time display in tooltips
- Fix re-auth button to use correct CCS endpoint
- Reduce quota cache staleness (30s stale, 1m refresh)
CLIProxyAPI intentionally doesn't persist refreshed tokens to disk
(to prevent refresh loops), so file-based expiry was misleading.
Dashboard now shows truthful operational state from runtime stats.
- Add quota-fetcher.ts for Google Cloud Code API integration
- Add REST endpoint GET /api/cliproxy/quota/:provider/:accountId
- Add useAccountQuota hook with React Query caching
- Display quota bar in AccountItem with per-model tooltip
- Create reusable Progress component
- Consolidate quota types in api-client.ts
- Wrap JSON.parse in try-catch with clear error message for malformed JSON
- Add iflow, kiro, ghcp providers to CLIProxyVariantConfig type
- Make settings optional in CLIProxyVariantConfig (was required with empty string fallback)
- Remove unsafe type cast in loadConfigSafe()
Addresses additional edge cases from PR #215 code review.
Recovery failures (permission errors, disk full) no longer block
--version/--help commands. Recovery is best-effort - warns on failure
but allows basic CLI functionality to continue.
Fixes edge case identified in PR #215 code review.
Fixes#214 - Fresh install fails with 'Config not found' when running
ccs config before RecoveryManager runs.
Changes:
- Move RecoveryManager.recoverAll() before all early-exit commands in ccs.ts
- Fix loadConfigSafe() to return empty config instead of throwing in legacy mode
- Add getActiveConfigPath() for mode-aware config path resolution
- Rename cliproxy getConfigPath() to getCliproxyConfigPath() to avoid confusion
- Update help-command.ts to show correct config path based on mode
This ensures all commands benefit from auto-recovery on fresh installs,
and gracefully handles missing config files in all code paths.
When Kiro OAuth callback redirects to Kiro IDE instead of CLI, this fix:
- Auto-attempts token import from Kiro IDE storage
- Adds --import flag for manual import (ccs kiro --import)
- Shows clear instructions if import fails
Fixes#212
The try/catch around loadConfig() in getSettingsPath() was ineffective
because process.exit() cannot be caught by try/catch - it terminates
the process immediately.
Replace with inline safe logic that:
1. Checks if config.json exists with fs.existsSync()
2. Reads and parses JSON manually
3. Uses isConfig() type guard for validation
4. Properly catches JSON parse errors
This ensures unified mode users don't crash when falling back to check
legacy config.json for profiles not found in config.yaml.
Add loadConfigSafe() function that handles both unified (config.yaml) and
legacy (config.json) config formats. Uses throwable errors instead of
process.exit() so try/catch blocks work properly in web server routes.
Updated files to use loadConfigSafe():
- overview-routes.ts: Dashboard overview API
- route-helpers.ts: readConfigSafe() helper
- profile-reader.ts: API profile reading
- profile-writer.ts: API profile writing
- variant-config-adapter.ts: CLIProxy variant config
Fixes#206 (Problem 2: config.json not found when user has config.yaml)
- Add SIGKILL escalation after 3s SIGTERM timeout in session-tracker
- Auto-stop running proxy before binary update in binary-manager
- Add waitForPortFree utility to port-utils
- Detect version mismatch on startup and auto-restart outdated proxy
- Store and expose version in session lock for detection
Fixes issue where old proxy version continues running after update,
causing UI to show different version than actually running.
Linux kernel truncates process names to 15 chars, causing
'cli-proxy-api-plus' to appear as 'cli-proxy-api-p' in lsof output.
Switch from exact whitelist to prefix matching for robust detection.
- Check config.yaml first (v2 format), fallback to config.json (legacy)
- Make glm.settings.json and kimi.settings.json optional
- Only validate settings files if they exist
- Fix false positives: postinstall no longer creates these files
Fixes health check reporting errors for missing files that are now
optional and created on-demand when user configures a profile.
Two bugs caused the health check to incorrectly report "Proxy not running":
1. lsof command used mutually exclusive flags (-t and -F)
- Removed -t flag since -F already provides structured output
2. isCLIProxyProcess whitelist was missing common binary names
- Added: cliproxyapi, cli-proxy-api-plus (and .exe variants)
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Reduce config.yaml header from 26 lines to 2 lines
- Replace auto-save on blur with explicit Save button in Auth section
- Add hasChanges detection to enable/disable Save button
- Remove unused saveApiKey and saveSecret functions
- GET /api/settings/auth/tokens - masked token status
- GET /api/settings/auth/tokens/raw - unmasked (Cache-Control: no-store)
- PUT /api/settings/auth/tokens - update API key or management secret
- POST /api/settings/auth/tokens/regenerate-secret - generate secure secret
- POST /api/settings/auth/tokens/reset - reset to defaults
- overview-routes: Use ProfileRegistry for account count, merge both
legacy and unified sources
- file-watcher: Add config.yaml to watch list for real-time updates
in unified mode
Related to #203
Dashboard accounts page was only reading from profiles.json, ignoring
accounts stored in config.yaml when in unified mode. This caused
accounts created via CLI to be invisible in the dashboard.
- Use ProfileRegistry instead of direct file access
- Merge legacy and unified accounts (unified takes precedence)
- Set default via unified config when in unified mode
Fixes#203
Call initUI() at the start of main() to ensure chalk is loaded
before any status messages. Previously, colors only worked when
an update notification was displayed because showUpdateNotification()
was the only place calling initUI().
Fixes#201
When CCS updates trigger CLIProxy config regeneration, user-added API
keys in ~/.ccs/cliproxy/config.yaml were being overwritten with only
the internal CCS key.
Changes:
- Add parseUserApiKeys() to extract user keys from existing config
- Modify regenerateConfig() to preserve user API keys alongside port
- Update generateUnifiedConfigContent() to accept userApiKeys param
- Add comprehensive test suite (14 tests) for key preservation
Fixes#200
Add three new Anthropic-compatible API providers as presets:
- Minimax: M2.1/M2.1-lightning/M2 models with 1M context
- DeepSeek: V3.2 and R1 reasoning model (128K context)
- Qwen: Alibaba Cloud qwen3-coder-plus (256K context)
Closes#123
- variant-settings.ts: Use expandPath() in deleteSettingsFile()
- profile-writer.ts: Use expandPath() in removeApiProfileUnified()
- migration-manager.ts: Remove local expandPath(), import from helpers
All path expansion now uses the central expandPath() from utils/helpers.ts
which handles ~, ~\, ${VAR}, $VAR, and %VAR% on all platforms.
- Check legacy config.json/profiles.json in isFirstTimeInstall()
- Use expandPath() for cross-platform path handling in profile-detector
- Add pre-flight API key validation for better error messages
- Enhance 401 error handling with Z.AI refresh guidance
Fixes#195
- Change default to noIncognito=true (normal browser) for reliability
- Always print OAuth URL to terminal for VS Code popup detection
- Move incognito toggle from Proxy settings to Kiro provider page
- Add --incognito flag to opt into incognito mode (was --no-incognito)
- Update help text to reflect new defaults
- Add "Kiro: Use normal browser" toggle in config dashboard (Proxy section)
- Add --no-incognito documentation to ccs --help output
- Show hint after successful Kiro auth about saving AWS credentials
Kiro OAuth always opens in incognito browser mode, which doesn't save
login credentials. Users need to re-enter email each authentication.
Changes:
- Add noIncognito option to OAuthOptions interface
- Add kiro_no_incognito config option to CLIProxyConfig
- Pass --no-incognito flag to CLIProxyAPI binary when enabled
- Support both CLI flag (--no-incognito) and config.yaml setting
Usage:
- CLI: ccs kiro --auth --no-incognito
- Config: cliproxy.kiro_no_incognito: true in config.yaml
Enables running multiple CLIProxy variants simultaneously on different
ports. Each variant now gets a unique port in the 18100-18199 range,
allowing concurrent use of providers like Gemini + Codex + custom variants.
Key changes:
- Add port field to variant config schema
- Implement automatic port allocation (18100 + variant index)
- Support variant-specific settings paths in config generator
- Display port in dashboard UI for debugging
- Show all models in variant editor dropdown
- Add comprehensive tests for port allocation edge cases
Closes related variant isolation work.
Device Code flows (Qwen, GHCP) may require interactive terminal input
before generating the device code. For example, Qwen prompts for email.
Previously, all auth processes used piped stdin which blocked user input.
Now Device Code flows use 'inherit' for stdin, allowing users to respond
to prompts directly in the terminal.
Authorization Code flows (Gemini, Codex) still use piped stdin for
programmatic project selection.
Closes#188
- Fix port default logic: HTTP→8317, HTTPS→443 (was always 8317)
- Fix enabled:false being ignored in proxy config resolution
- Update stale comments referencing HTTP:80 to HTTP:8317
- Add 9 unit tests for edge cases (enabled handling, port defaults)
kaitranntt