llm-switcher 1.1.11 → 1.2.2

package/CHANGELOG.md

@@ -0,0 +1,63 @@
+# Changelog — LLM Switcher
+
+## Release 1.2.2
+
+- **Structured output reaches every upstream:** A request that asks for JSON that matches a schema now keeps that schema. Before this release, the gateway did not read `output_config.format` (Claude Code) or `text.format` (Codex), so the provider got a free-text request. The schema now goes to the provider as `response_format` (OpenAI Chat), `output_config.format` (Anthropic), or `responseMimeType` with `responseSchema` (Gemini and Vertex). A request for JSON without a schema reaches Anthropic as plain text, because Anthropic has no JSON mode without a schema.
+
+## Release 1.2.1
+
+- **Intact Visual Architecture & Hash Routing:** Fully restructured dashboard using Intact design system, pure CSS tokens, and hash-based client routing (`#/routes`, `#/models`, `#/logs`, `#/doctor`).
+- **Official Brand Icon Assets:** Integrated official brand icons for Anthropic Claude, OpenAI Codex, Google Gemini, Intact, OpenRouter, and Ollama, served statically under `/icons/*` with immutable caching and PNG/SVG whitelisting.
+- **Prefetched Navigation Counters:** Navigation badges prefetch model discovery and request inspector counts immediately on page load, eliminating the delay where badges showed zero until the tab was selected.
+- **Hardened Profile Slot Mapping:** Synchronized Claude model tiers (`sonnet`, `opus`, `haiku`, `fable`) and Codex model roles (`main`, `review`, `subagent`), ensuring slot persistence when editing profiles and removing redundant role tabs for Claude-targeted profiles.
+- **CORS & Endpoint Fixes:** Added `x-llm-switcher-token` to preflight `Access-Control-Allow-Headers` and resolved routing precedence for `GET /api/catalog`.
+
+## Release 1.2.0
+
+- **Zero-Mutation Interceptor Invariant:** The gateway now operates strictly in the network path via the blindfold interceptor, never modifying user configuration files like `~/.claude/settings.json` or `~/.codex/config.toml`. Loopback base URLs (`ANTHROPIC_BASE_URL` and `OPENAI_BASE_URL`) are scrubbed by shims so traffic routes through `HTTPS_PROXY` cleanly.
+- **Concurrent Multi-Tool Support (`claude` & `codex`):** `activeProfiles` in `config.json` now configures `{ claude, codex }` independently. Switching tools updates the interceptor dynamically over `POST /_control/active-tools` without dropping in-flight connections or restarting ports.
+- **Auto-Discovery & Dynamic Model Catalog:** Automatically discovers official models for Claude Code & Codex; detects tool version updates and refreshes mappings on the fly (`switch models`).
+- **Automated Configuration Migration:** Legacy configurations with single `activeProfile` or legacy format keys are atomically migrated on startup with automatic backup (`config.json.bak-*`) and CAS protection.
+- **Gemini / Vertex Schema Healing:** Empty tool input schemas (`{}`) are automatically normalized to valid JSON schema objects with type definition before forwarding upstream, eliminating HTTP 400 schema rejection errors.
+- **Windows Shim Reliability:** Windows `.cmd` launcher shims enforce strict CRLF formatting, preventing cmd.exe label lookup offsets (`SCRUB_TIER`), and reference the certificate bundle at `blindfold/certs/ca.pem`.
+- **Refreshed Web UI:** Modernized dashboard with real-time target indicators, tool-specific profiles, and model context-window tracking.
+
+## Release 1.1.10
+
+- **Self-improvement with intact:** Integration with intact credential proxy and contract lab.
+
+## Release 1.1.9
+
+- **Dashboard:** Open `http://127.0.0.1:3456/ui` directly. The page carries the admin token securely.
+
+## Release 1.1.8
+
+- **Claude Code tools:** Fixed tool schema values of `0`, `false`, `""` or `null` being converted to empty object schemas, resolving Gemini HTTP 400 errors.
+- **Shim PATH order:** Improved PATH order diagnostic in `switch shim status` and `switch doctor`.
+
+## Release 1.1.7
+
+- **Codex tools:** Model catalog keeps Codex in direct tool mode; reads tools from `additional_tools` input items.
+
+## Release 1.1.6
+
+- **Codex over WebSocket:** Gateway tracks WebSocket sessions. Multi-turn sessions using `previous_response_id` retrieve earlier turns seamlessly.
+- **Codex warmup:** `response.create` with `generate: false` answered locally without spending model calls.
+- **Codex model name:** Fixed false "high-risk cyber activity" warnings by stopping unwanted `OpenAI-Model: main` headers.
+
+## Release 1.1.5
+
+- **Contract lab privacy:** Masking allowlist ensures all client prompts, code, and files are completely redacted before uploading samples.
+
+## Release 1.1.4
+
+- **Contract lab privacy:** Initial client-side masking implementation.
+
+## Release 1.1.3
+
+- **Dashboard:** Improved access token error messaging and reorganized model settings.
+
+## Release 1.1.2
+
+- **npm package:** Released as global npm package `llm-switcher`.
+- **LibreSSL compatibility:** Added macOS default LibreSSL support for certificate generation.