chore: Bump version to 0.1.5

- Fix orphaned tool_use error by tracking skip state per call_id
- Handle retries with same call_id correctly
- Add comprehensive debug logging for troubleshooting

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>

.github/workflows/rust-release.yml

@@ -476,7 +476,7 @@ jobs:
           tag: ${{ github.ref_name }}
           config: .github/dotslash-config.json
 
-  # Publish to npm using authentication token
+  # Publish to npm using Trusted Publishers (OIDC)
   publish-npm:
     # Publish to npm for stable releases and alpha pre-releases with numeric suffixes.
     if: ${{ needs.release.outputs.should_publish_npm == 'true' }}
@@ -485,6 +485,7 @@ jobs:
     runs-on: ubuntu-latest
     permissions:
       contents: read
+      id-token: write  # Required for OIDC authentication
 
     steps:
       - name: Setup Node.js
@@ -492,6 +493,10 @@ jobs:
         with:
           node-version: 22
           registry-url: "https://registry.npmjs.org"
+          scope: "@valknarthing"
+
+      - name: Update npm
+        run: npm install -g npm@latest
 
       - name: Download npm tarballs from release
         env:
@@ -511,10 +516,6 @@ jobs:
           VERSION: ${{ needs.release.outputs.version }}
           NPM_TAG: ${{ needs.release.outputs.npm_tag }}
         run: |
-          # Write auth token to the .npmrc file that setup-node created
-          echo "//registry.npmjs.org/:_authToken=${{ secrets.NPM_TOKEN }}" >> ${NPM_CONFIG_USERCONFIG}
-
-
           set -euo pipefail
           tag_args=()
           if [[ -n "${NPM_TAG}" ]]; then
@@ -526,24 +527,24 @@ jobs:
           )
 
           for tarball in "${tarballs[@]}"; do
-            npm publish "${GITHUB_WORKSPACE}/dist/npm/${tarball}" --access public "${tag_args[@]}"
+            npm publish "${GITHUB_WORKSPACE}/dist/npm/${tarball}" --provenance --access public "${tag_args[@]}"
           done
 
-  update-branch:
+  # update-branch:
-    name: Update latest-alpha-cli branch
+  #   name: Update latest-alpha-cli branch
-    permissions:
+  #   permissions:
-      contents: write
+  #     contents: write
-    needs: release
+  #   needs: release
-    runs-on: ubuntu-latest
+  #   runs-on: ubuntu-latest
-
+  #
-    steps:
+  #   steps:
-      - name: Update latest-alpha-cli branch
+  #     - name: Update latest-alpha-cli branch
-        env:
+  #       env:
-          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+  #         GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-        run: |
+  #       run: |
-          set -euo pipefail
+  #         set -euo pipefail
-          gh api \
+  #         gh api \
-            repos/${GITHUB_REPOSITORY}/git/refs/heads/latest-alpha-cli \
+  #           repos/${GITHUB_REPOSITORY}/git/refs/heads/latest-alpha-cli \
-            -X PATCH \
+  #           -X PATCH \
-            -f sha="${GITHUB_SHA}" \
+  #           -f sha="${GITHUB_SHA}" \
-            -F force=true
+  #           -F force=true

README.md

@@ -47,56 +47,56 @@ LLMX is powered by [LiteLLM](https://docs.litellm.ai/), which provides access to
 
 ```bash
 # Set your LiteLLM server URL (default: http://localhost:4000/v1)
-export LITELLM_BASE_URL="http://localhost:4000/v1"
+export LLMX_BASE_URL="http://localhost:4000/v1"
-export LITELLM_API_KEY="your-api-key"
+export LLMX_API_KEY="your-api-key"
 
 # Run LLMX
 llmx "hello world"
 ```
 
-**Configuration:** See [LITELLM-SETUP.md](./LITELLM-SETUP.md) for detailed setup instructions.
+**Configuration:** See [LITELLM-SETUP.md](https://github.com/valknarthing/llmx/blob/main/LITELLM-SETUP.md) for detailed setup instructions.
 
-You can also use LLMX with ChatGPT or OpenAI API keys. For authentication options, see the [authentication docs](./docs/authentication.md).
+You can also use LLMX with ChatGPT or OpenAI API keys. For authentication options, see the [authentication docs](https://github.com/valknarthing/llmx/blob/main/docs/authentication.md).
 
 ### Model Context Protocol (MCP)
 
-LLMX can access MCP servers. To configure them, refer to the [config docs](./docs/config.md#mcp_servers).
+LLMX can access MCP servers. To configure them, refer to the [config docs](https://github.com/valknarthing/llmx/blob/main/docs/config.md#mcp_servers).
 
 ### Configuration
 
-LLMX CLI supports a rich set of configuration options, with preferences stored in `~/.llmx/config.toml`. For full configuration options, see [Configuration](./docs/config.md).
+LLMX CLI supports a rich set of configuration options, with preferences stored in `~/.llmx/config.toml`. For full configuration options, see [Configuration](https://github.com/valknarthing/llmx/blob/main/docs/config.md).
 
 ---
 
 ### Docs & FAQ
 
-- [**Getting started**](./docs/getting-started.md)
+- [**Getting started**](https://github.com/valknarthing/llmx/blob/main/docs/getting-started.md)
-  - [CLI usage](./docs/getting-started.md#cli-usage)
+  - [CLI usage](https://github.com/valknarthing/llmx/blob/main/docs/getting-started.md#cli-usage)
-  - [Slash Commands](./docs/slash_commands.md)
+  - [Slash Commands](https://github.com/valknarthing/llmx/blob/main/docs/slash_commands.md)
-  - [Running with a prompt as input](./docs/getting-started.md#running-with-a-prompt-as-input)
+  - [Running with a prompt as input](https://github.com/valknarthing/llmx/blob/main/docs/getting-started.md#running-with-a-prompt-as-input)
-  - [Example prompts](./docs/getting-started.md#example-prompts)
+  - [Example prompts](https://github.com/valknarthing/llmx/blob/main/docs/getting-started.md#example-prompts)
-  - [Custom prompts](./docs/prompts.md)
+  - [Custom prompts](https://github.com/valknarthing/llmx/blob/main/docs/prompts.md)
-  - [Memory with AGENTS.md](./docs/getting-started.md#memory-with-agentsmd)
+  - [Memory with AGENTS.md](https://github.com/valknarthing/llmx/blob/main/docs/getting-started.md#memory-with-agentsmd)
-- [**Configuration**](./docs/config.md)
+- [**Configuration**](https://github.com/valknarthing/llmx/blob/main/docs/config.md)
-  - [Example config](./docs/example-config.md)
+  - [Example config](https://github.com/valknarthing/llmx/blob/main/docs/example-config.md)
-- [**Sandbox & approvals**](./docs/sandbox.md)
+- [**Sandbox & approvals**](https://github.com/valknarthing/llmx/blob/main/docs/sandbox.md)
-- [**Authentication**](./docs/authentication.md)
+- [**Authentication**](https://github.com/valknarthing/llmx/blob/main/docs/authentication.md)
-  - [Auth methods](./docs/authentication.md#forcing-a-specific-auth-method-advanced)
+  - [Auth methods](https://github.com/valknarthing/llmx/blob/main/docs/authentication.md#forcing-a-specific-auth-method-advanced)
-  - [Login on a "Headless" machine](./docs/authentication.md#connecting-on-a-headless-machine)
+  - [Login on a "Headless" machine](https://github.com/valknarthing/llmx/blob/main/docs/authentication.md#connecting-on-a-headless-machine)
 - **Automating LLMX**
   - [GitHub Action](https://github.com/valknarthing/llmx-action)
-  - [TypeScript SDK](./sdk/typescript/README.md)
+  - [TypeScript SDK](https://github.com/valknarthing/llmx/blob/main/sdk/typescript/README.md)
-  - [Non-interactive mode (`llmx exec`)](./docs/exec.md)
+  - [Non-interactive mode (`llmx exec`)](https://github.com/valknarthing/llmx/blob/main/docs/exec.md)
-- [**Advanced**](./docs/advanced.md)
+- [**Advanced**](https://github.com/valknarthing/llmx/blob/main/docs/advanced.md)
-  - [Tracing / verbose logging](./docs/advanced.md#tracing--verbose-logging)
+  - [Tracing / verbose logging](https://github.com/valknarthing/llmx/blob/main/docs/advanced.md#tracing--verbose-logging)
-  - [Model Context Protocol (MCP)](./docs/advanced.md#model-context-protocol-mcp)
+  - [Model Context Protocol (MCP)](https://github.com/valknarthing/llmx/blob/main/docs/advanced.md#model-context-protocol-mcp)
-- [**Zero data retention (ZDR)**](./docs/zdr.md)
+- [**Zero data retention (ZDR)**](https://github.com/valknarthing/llmx/blob/main/docs/zdr.md)
-- [**Contributing**](./docs/contributing.md)
+- [**Contributing**](https://github.com/valknarthing/llmx/blob/main/docs/contributing.md)
-- [**Install & build**](./docs/install.md)
+- [**Install & build**](https://github.com/valknarthing/llmx/blob/main/docs/install.md)
-  - [System Requirements](./docs/install.md#system-requirements)
+  - [System Requirements](https://github.com/valknarthing/llmx/blob/main/docs/install.md#system-requirements)
-  - [DotSlash](./docs/install.md#dotslash)
+  - [DotSlash](https://github.com/valknarthing/llmx/blob/main/docs/install.md#dotslash)
-  - [Build from source](./docs/install.md#build-from-source)
+  - [Build from source](https://github.com/valknarthing/llmx/blob/main/docs/install.md#build-from-source)
-- [**FAQ**](./docs/faq.md)
+- [**FAQ**](https://github.com/valknarthing/llmx/blob/main/docs/faq.md)
 
 ---
 

llmx-cli/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@valknarthing/llmx",
-  "version": "0.1.1",
+  "version": "0.1.2",
   "license": "Apache-2.0",
   "description": "LLMX CLI - Multi-provider coding agent powered by LiteLLM",
   "bin": {

llmx-rs/Cargo.lock

@@ -178,7 +178,7 @@ checksum = "a23eb6b1614318a8071c9b2521f36b424b2c83db5eb3a0fead4a6c0809af6e61"
 
 [[package]]
 name = "app_test_support"
-version = "0.1.1"
+version = "0.1.5"
 dependencies = [
  "anyhow",
  "assert_cmd",
@@ -945,7 +945,7 @@ checksum = "773648b94d0e5d620f64f280777445740e61fe701025087ec8b57f45c791888b"
 
 [[package]]
 name = "core_test_support"
-version = "0.1.1"
+version = "0.1.5"
 dependencies = [
  "anyhow",
  "assert_cmd",
@@ -1192,7 +1192,7 @@ dependencies = [
 
 [[package]]
 name = "deadpool-runtime"
-version = "0.1.4"
+version = "0.1.5"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "092966b41edc516079bdf31ec78a2e0588d1d0c08f78b91d8307215928642b2b"
 
@@ -2501,7 +2501,7 @@ dependencies = [
 
 [[package]]
 name = "inout"
-version = "0.1.4"
+version = "0.1.5"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "879f10e63c20629ecabbb64a8010319738c66a5cd0c29b02d63d272b03751d01"
 dependencies = [
@@ -2822,7 +2822,7 @@ checksum = "241eaef5fd12c88705a01fc1066c48c4b36e0dd4377dcdc7ec3942cea7a69956"
 
 [[package]]
 name = "llmx-ansi-escape"
-version = "0.1.1"
+version = "0.1.5"
 dependencies = [
  "ansi-to-tui",
  "ratatui",
@@ -2831,7 +2831,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-app-server"
-version = "0.1.1"
+version = "0.1.5"
 dependencies = [
  "anyhow",
  "app_test_support",
@@ -2866,7 +2866,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-app-server-protocol"
-version = "0.1.1"
+version = "0.1.5"
 dependencies = [
  "anyhow",
  "clap",
@@ -2884,7 +2884,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-apply-patch"
-version = "0.1.1"
+version = "0.1.5"
 dependencies = [
  "anyhow",
  "assert_cmd",
@@ -2899,7 +2899,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-arg0"
-version = "0.1.1"
+version = "0.1.5"
 dependencies = [
  "anyhow",
  "dotenvy",
@@ -2912,7 +2912,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-async-utils"
-version = "0.1.1"
+version = "0.1.5"
 dependencies = [
  "async-trait",
  "pretty_assertions",
@@ -2936,7 +2936,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-backend-openapi-models"
-version = "0.1.1"
+version = "0.1.5"
 dependencies = [
  "serde",
  "serde_json",
@@ -2945,7 +2945,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-chatgpt"
-version = "0.1.1"
+version = "0.1.5"
 dependencies = [
  "anyhow",
  "clap",
@@ -2960,7 +2960,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-cli"
-version = "0.1.1"
+version = "0.1.5"
 dependencies = [
  "anyhow",
  "assert_cmd",
@@ -3000,7 +3000,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-cloud-tasks"
-version = "0.1.1"
+version = "0.1.5"
 dependencies = [
  "anyhow",
  "async-trait",
@@ -3026,7 +3026,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-cloud-tasks-client"
-version = "0.1.1"
+version = "0.1.5"
 dependencies = [
  "anyhow",
  "async-trait",
@@ -3041,7 +3041,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-common"
-version = "0.1.1"
+version = "0.1.5"
 dependencies = [
  "clap",
  "llmx-app-server-protocol",
@@ -3053,7 +3053,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-core"
-version = "0.1.1"
+version = "0.1.5"
 dependencies = [
  "anyhow",
  "askama",
@@ -3134,7 +3134,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-exec"
-version = "0.1.1"
+version = "0.1.5"
 dependencies = [
  "anyhow",
  "assert_cmd",
@@ -3167,7 +3167,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-execpolicy"
-version = "0.1.1"
+version = "0.1.5"
 dependencies = [
  "allocative",
  "anyhow",
@@ -3187,7 +3187,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-feedback"
-version = "0.1.1"
+version = "0.1.5"
 dependencies = [
  "anyhow",
  "llmx-protocol",
@@ -3198,7 +3198,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-file-search"
-version = "0.1.1"
+version = "0.1.5"
 dependencies = [
  "anyhow",
  "clap",
@@ -3211,7 +3211,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-git"
-version = "0.1.1"
+version = "0.1.5"
 dependencies = [
  "assert_matches",
  "once_cell",
@@ -3227,7 +3227,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-keyring-store"
-version = "0.1.1"
+version = "0.1.5"
 dependencies = [
  "keyring",
  "tracing",
@@ -3235,7 +3235,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-linux-sandbox"
-version = "0.1.1"
+version = "0.1.5"
 dependencies = [
  "clap",
  "landlock",
@@ -3248,7 +3248,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-login"
-version = "0.1.1"
+version = "0.1.5"
 dependencies = [
  "anyhow",
  "base64",
@@ -3272,7 +3272,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-mcp-server"
-version = "0.1.1"
+version = "0.1.5"
 dependencies = [
  "anyhow",
  "assert_cmd",
@@ -3299,7 +3299,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-ollama"
-version = "0.1.1"
+version = "0.1.5"
 dependencies = [
  "assert_matches",
  "async-stream",
@@ -3315,7 +3315,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-otel"
-version = "0.1.1"
+version = "0.1.5"
 dependencies = [
  "chrono",
  "eventsource-stream",
@@ -3336,14 +3336,14 @@ dependencies = [
 
 [[package]]
 name = "llmx-process-hardening"
-version = "0.1.1"
+version = "0.1.5"
 dependencies = [
  "libc",
 ]
 
 [[package]]
 name = "llmx-protocol"
-version = "0.1.1"
+version = "0.1.5"
 dependencies = [
  "anyhow",
  "base64",
@@ -3369,7 +3369,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-responses-api-proxy"
-version = "0.1.1"
+version = "0.1.5"
 dependencies = [
  "anyhow",
  "clap",
@@ -3385,7 +3385,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-rmcp-client"
-version = "0.1.1"
+version = "0.1.5"
 dependencies = [
  "anyhow",
  "axum",
@@ -3414,7 +3414,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-stdio-to-uds"
-version = "0.1.1"
+version = "0.1.5"
 dependencies = [
  "anyhow",
  "assert_cmd",
@@ -3425,7 +3425,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-tui"
-version = "0.1.1"
+version = "0.1.5"
 dependencies = [
  "anyhow",
  "arboard",
@@ -3490,7 +3490,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-utils-cache"
-version = "0.1.1"
+version = "0.1.5"
 dependencies = [
  "lru",
  "sha1",
@@ -3499,7 +3499,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-utils-image"
-version = "0.1.1"
+version = "0.1.5"
 dependencies = [
  "base64",
  "image",
@@ -3511,7 +3511,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-utils-json-to-toml"
-version = "0.1.1"
+version = "0.1.5"
 dependencies = [
  "pretty_assertions",
  "serde_json",
@@ -3520,7 +3520,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-utils-pty"
-version = "0.1.1"
+version = "0.1.5"
 dependencies = [
  "anyhow",
  "portable-pty",
@@ -3529,7 +3529,7 @@ dependencies = [
 
 [[package]]
 name = "llmx-utils-readiness"
-version = "0.1.1"
+version = "0.1.5"
 dependencies = [
  "assert_matches",
  "async-trait",
@@ -3540,11 +3540,11 @@ dependencies = [
 
 [[package]]
 name = "llmx-utils-string"
-version = "0.1.1"
+version = "0.1.5"
 
 [[package]]
 name = "llmx-utils-tokenizer"
-version = "0.1.1"
+version = "0.1.5"
 dependencies = [
  "anyhow",
  "pretty_assertions",
@@ -3655,7 +3655,7 @@ checksum = "47e1ffaa40ddd1f3ed91f717a33c8c0ee23fff369e3aa8772b9605cc1d22f4c3"
 
 [[package]]
 name = "mcp-types"
-version = "0.1.1"
+version = "0.1.5"
 dependencies = [
  "schemars 0.8.22",
  "serde",
@@ -3665,7 +3665,7 @@ dependencies = [
 
 [[package]]
 name = "mcp_test_support"
-version = "0.1.1"
+version = "0.1.5"
 dependencies = [
  "anyhow",
  "assert_cmd",
@@ -3942,7 +3942,7 @@ checksum = "51d515d32fb182ee37cda2ccdcb92950d6a3c2893aa280e540671c2cd0f3b1d9"
 
 [[package]]
 name = "num-integer"
-version = "0.1.46"
+version = "0.1.56"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "7969661fd2958a5cb096e56c8e1ad0444ac2bbcd0061bd28660485a44879858f"
 dependencies = [
@@ -3951,7 +3951,7 @@ dependencies = [
 
 [[package]]
 name = "num-iter"
-version = "0.1.45"
+version = "0.1.55"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "1429034a0490724d0075ebb2bc9e875d6503c3cf69e235a8941aa757d83ef5bf"
 dependencies = [
@@ -4520,7 +4520,7 @@ dependencies = [
 
 [[package]]
 name = "potential_utf"
-version = "0.1.4"
+version = "0.1.5"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "b73949432f5e2a09657003c25bca5e19a0e9c84f8058ca374f49e0ebe605af77"
 dependencies = [
@@ -6580,7 +6580,7 @@ checksum = "8df9b6e13f2d32c91b9bd719c00d1958837bc7dec474d94952798cc8e69eeec3"
 
 [[package]]
 name = "tracing"
-version = "0.1.41"
+version = "0.1.51"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "784e0ac535deb450455cbfa28a6f0df145ea1bb7ae51b821cf5e7927fdcfbdd0"
 dependencies = [

llmx-rs/Cargo.toml

@@ -43,7 +43,7 @@ members = [
 resolver = "2"
 
 [workspace.package]
-version = "0.1.1"
+version = "0.1.5"
 # Track the edition for all workspace crates in one place. Individual
 # crates can still override this value, but keeping it here means new
 # crates created with `cargo new -w ...` automatically inherit the 2024
@@ -191,7 +191,7 @@ tokio-util = "0.7.16"
 toml = "0.9.5"
 toml_edit = "0.23.4"
 tonic = "0.13.1"
-tracing = "0.1.41"
+tracing = "0.1.51"
 tracing-appender = "0.2.3"
 tracing-subscriber = "0.3.20"
 tracing-test = "0.2.5"

llmx-rs/app-server/tests/common/mcp_process.rs

@@ -138,7 +138,7 @@ impl McpProcess {
             client_info: ClientInfo {
                 name: "llmx-app-server-tests".to_string(),
                 title: None,
-                version: "0.1.1".to_string(),
+                version: "0.1.5".to_string(),
             },
         })?);
         let req_id = self.send_request("initialize", params).await?;

llmx-rs/app-server/tests/suite/user_agent.rs

@@ -26,7 +26,7 @@ async fn get_user_agent_returns_current_llmx_user_agent() -> Result<()> {
 
     let os_info = os_info::get();
     let user_agent = format!(
-        "llmx_cli_rs/0.1.1 ({} {}; {}) {} (llmx-app-server-tests; 0.1.1)",
+        "llmx_cli_rs/0.1.5 ({} {}; {}) {} (llmx-app-server-tests; 0.1.5)",
         os_info.os_type(),
         os_info.version(),
         os_info.architecture().unwrap_or("unknown"),

llmx-rs/core/src/chat_completions.rs

@@ -161,7 +161,65 @@ pub(crate) async fn stream_chat_completions(
     // aggregated assistant message was recorded alongside an earlier partial).
     let mut last_assistant_text: Option<String> = None;
 
+    // Build a map of which call_ids have outputs
+    // We'll use this to ensure we never send a FunctionCall without its corresponding output
+    let mut call_ids_with_outputs: std::collections::HashSet<String> = std::collections::HashSet::new();
+
+    // First pass: collect all call_ids that have outputs
+    for item in input.iter() {
+        if let ResponseItem::FunctionCallOutput { call_id, .. } = item {
+            call_ids_with_outputs.insert(call_id.clone());
+        }
+    }
+
+    debug!("=== Chat Completions Request Debug ===");
+    debug!("Input items count: {}", input.len());
+    debug!("Call IDs with outputs: {:?}", call_ids_with_outputs);
+
+    // Second pass: find the first FunctionCall that doesn't have an output
+    let mut cutoff_at_idx: Option<usize> = None;
     for (idx, item) in input.iter().enumerate() {
+        if let ResponseItem::FunctionCall { call_id, name, .. } = item {
+            if !call_ids_with_outputs.contains(call_id) {
+                debug!("Found unanswered function call '{}' (call_id: {}) at index {}", name, call_id, idx);
+                cutoff_at_idx = Some(idx);
+                break;
+            }
+        }
+    }
+
+    if let Some(cutoff) = cutoff_at_idx {
+        debug!("Cutting off at index {} to avoid orphaned tool calls", cutoff);
+    } else {
+        debug!("No unanswered function calls found, processing all items");
+    }
+
+    // Track whether the MOST RECENT FunctionCall with each call_id was skipped
+    // This allows the same call_id to be retried - we only skip outputs for the specific skipped calls
+    let mut call_id_skip_state: std::collections::HashMap<String, bool> = std::collections::HashMap::new();
+
+    for (idx, item) in input.iter().enumerate() {
+        // Stop processing if we've reached an unanswered function call
+        if let Some(cutoff) = cutoff_at_idx {
+            if idx >= cutoff {
+                debug!("Stopping at index {} due to unanswered function call", idx);
+                break;
+            }
+        }
+
+        debug!("Processing item {} of type: {}", idx, match item {
+            ResponseItem::Message { role, .. } => format!("Message(role={})", role),
+            ResponseItem::FunctionCall { name, call_id, .. } => format!("FunctionCall(name={}, call_id={})", name, call_id),
+            ResponseItem::FunctionCallOutput { call_id, .. } => format!("FunctionCallOutput(call_id={})", call_id),
+            ResponseItem::LocalShellCall { .. } => "LocalShellCall".to_string(),
+            ResponseItem::CustomToolCall { .. } => "CustomToolCall".to_string(),
+            ResponseItem::CustomToolCallOutput { .. } => "CustomToolCallOutput".to_string(),
+            ResponseItem::Reasoning { .. } => "Reasoning".to_string(),
+            ResponseItem::WebSearchCall { .. } => "WebSearchCall".to_string(),
+            ResponseItem::GhostSnapshot { .. } => "GhostSnapshot".to_string(),
+            ResponseItem::Other => "Other".to_string(),
+        });
+
         match item {
             ResponseItem::Message { role, content, .. } => {
                 // Build content either as a plain string (typical for assistant text)
@@ -175,7 +233,10 @@ pub(crate) async fn stream_chat_completions(
                         ContentItem::InputText { text: t }
                         | ContentItem::OutputText { text: t } => {
                             text.push_str(t);
-                            items.push(json!({"type":"text","text": t}));
+                            // Only add text content blocks that are non-empty
+                            if !t.trim().is_empty() {
+                                items.push(json!({"type":"text","text": t}));
+                            }
                         }
                         ContentItem::InputImage { image_url } => {
                             saw_image = true;
@@ -184,6 +245,11 @@ pub(crate) async fn stream_chat_completions(
                     }
                 }
 
+                // Skip messages with empty or whitespace-only text content (unless they contain images)
+                if text.trim().is_empty() && !saw_image {
+                    continue;
+                }
+
                 // Skip exact-duplicate assistant messages.
                 if role == "assistant" {
                     if let Some(prev) = &last_assistant_text
@@ -219,6 +285,18 @@ pub(crate) async fn stream_chat_completions(
                 call_id,
                 ..
             } => {
+                // Validate that arguments is valid JSON before sending to API
+                // If invalid, skip this function call to avoid API errors
+                if serde_json::from_str::<serde_json::Value>(arguments).is_err() {
+                    debug!("Skipping malformed function call with invalid JSON arguments: {}", arguments);
+                    // Mark this call_id's most recent state as skipped
+                    call_id_skip_state.insert(call_id.clone(), true);
+                    continue;
+                }
+
+                // Mark this call_id's most recent state as NOT skipped (valid call)
+                call_id_skip_state.insert(call_id.clone(), false);
+
                 let mut msg = json!({
                     "role": "assistant",
                     "content": null,
@@ -263,6 +341,12 @@ pub(crate) async fn stream_chat_completions(
                 messages.push(msg);
             }
             ResponseItem::FunctionCallOutput { call_id, output } => {
+                // Skip outputs only if the MOST RECENT FunctionCall with this call_id was skipped
+                if call_id_skip_state.get(call_id) == Some(&true) {
+                    debug!("Skipping function call output for most recent skipped call_id: {}", call_id);
+                    continue;
+                }
+
                 // Prefer structured content items when available (e.g., images)
                 // otherwise fall back to the legacy plain-string content.
                 let content_value = if let Some(items) = &output.content_items {
@@ -328,14 +412,23 @@ pub(crate) async fn stream_chat_completions(
         }
     }
 
+    debug!("Built {} messages for API request", messages.len());
+    debug!("=== End Chat Completions Request Debug ===");
+
     let tools_json = create_tools_json_for_chat_completions_api(&prompt.tools)?;
-    let payload = json!({
+    let mut payload = json!({
         "model": model_family.slug,
         "messages": messages,
         "stream": true,
         "tools": tools_json,
     });
 
+    // Add max_tokens - required by Anthropic Messages API
+    // Use a sensible default of 8192 if not configured
+    if let Some(obj) = payload.as_object_mut() {
+        obj.insert("max_tokens".to_string(), json!(8192));
+    }
+
     debug!(
         "POST to {}: {}",
         provider.get_full_url(&None),

llmx-rs/core/src/tools/spec.rs

@@ -693,7 +693,7 @@ pub(crate) fn create_tools_json_for_chat_completions_api(
     // We start with the JSON for the Responses API and than rewrite it to match
     // the chat completions tool call format.
     let responses_api_tools_json = create_tools_json_for_responses_api(tools)?;
-    let tools_json = responses_api_tools_json
+    let mut tools_json = responses_api_tools_json
         .into_iter()
         .filter_map(|mut tool| {
             if tool.get("type") != Some(&serde_json::Value::String("function".to_string())) {
@@ -712,6 +712,14 @@ pub(crate) fn create_tools_json_for_chat_completions_api(
             }
         })
         .collect::<Vec<serde_json::Value>>();
+
+    // Add cache_control to the last tool to enable Anthropic prompt caching
+    if let Some(last_tool) = tools_json.last_mut() {
+        if let Some(obj) = last_tool.as_object_mut() {
+            obj.insert("cache_control".to_string(), json!({"type": "ephemeral"}));
+        }
+    }
+
     Ok(tools_json)
 }
 

llmx-rs/mcp-server/tests/common/mcp_process.rs

@@ -144,7 +144,7 @@ impl McpProcess {
         let initialized = self.read_jsonrpc_message().await?;
         let os_info = os_info::get();
         let user_agent = format!(
-            "llmx_cli_rs/0.1.1 ({} {}; {}) {} (elicitation test; 0.0.0)",
+            "llmx_cli_rs/0.1.5 ({} {}; {}) {} (elicitation test; 0.0.0)",
             os_info.os_type(),
             os_info.version(),
             os_info.architecture().unwrap_or("unknown"),
@@ -163,7 +163,7 @@ impl McpProcess {
                     "serverInfo": {
                         "name": "llmx-mcp-server",
                         "title": "LLMX",
-                        "version": "0.1.1",
+                        "version": "0.1.5",
                         "user_agent": user_agent
                     },
                     "protocolVersion": mcp_types::MCP_SCHEMA_VERSION

llmx-rs/tui/src/status/snapshots/llmx_tui__status__tests__status_snapshot_includes_monthly_limit.snap

@@ -5,7 +5,7 @@ expression: sanitized
 /status
 
 ╭───────────────────────────────────────────────────────────────────────────╮
-│  >_ LLMX (v0.1.1)                                                         │
+│  >_ LLMX (v0.1.5)                                                         │
 │                                                                           │
 │ Visit https://chatgpt.com/llmx/settings/usage for up-to-date              │
 │ information on rate limits and credits                                    │

llmx-rs/tui/src/status/snapshots/llmx_tui__status__tests__status_snapshot_includes_reasoning_details.snap

@@ -5,7 +5,7 @@ expression: sanitized
 /status
 
 ╭─────────────────────────────────────────────────────────────────╮
-│  >_ LLMX (v0.1.1)                                               │
+│  >_ LLMX (v0.1.5)                                               │
 │                                                                 │
 │ Visit https://chatgpt.com/llmx/settings/usage for up-to-date    │
 │ information on rate limits and credits                          │

llmx-rs/tui/src/status/snapshots/llmx_tui__status__tests__status_snapshot_shows_empty_limits_message.snap

@@ -5,7 +5,7 @@ expression: sanitized
 /status
 
 ╭──────────────────────────────────────────────────────────────╮
-│  >_ LLMX (v0.1.1)                                            │
+│  >_ LLMX (v0.1.5)                                            │
 │                                                              │
 │ Visit https://chatgpt.com/llmx/settings/usage for up-to-date │
 │ information on rate limits and credits                       │

llmx-rs/tui/src/status/snapshots/llmx_tui__status__tests__status_snapshot_shows_missing_limits_message.snap

@@ -5,7 +5,7 @@ expression: sanitized
 /status
 
 ╭──────────────────────────────────────────────────────────────╮
-│  >_ LLMX (v0.1.1)                                            │
+│  >_ LLMX (v0.1.5)                                            │
 │                                                              │
 │ Visit https://chatgpt.com/llmx/settings/usage for up-to-date │
 │ information on rate limits and credits                       │

llmx-rs/tui/src/status/snapshots/llmx_tui__status__tests__status_snapshot_shows_stale_limits_message.snap

@@ -5,7 +5,7 @@ expression: sanitized
 /status
 
 ╭───────────────────────────────────────────────────────────────────╮
-│  >_ LLMX (v0.1.1)                                                 │
+│  >_ LLMX (v0.1.5)                                                 │
 │                                                                   │
 │ Visit https://chatgpt.com/llmx/settings/usage for up-to-date      │
 │ information on rate limits and credits                            │

llmx-rs/tui/src/status/snapshots/llmx_tui__status__tests__status_snapshot_truncates_in_narrow_terminal.snap

@@ -5,7 +5,7 @@ expression: sanitized
 /status
 
 ╭────────────────────────────────────────────╮
-│  >_ LLMX (v0.1.1)                          │
+│  >_ LLMX (v0.1.5)                          │
 │                                            │
 │ Visit https://chatgpt.com/llmx/settings/   │
 │ usage for up-to-date                       │