Fix: Pass GitHub token to MCP servers in agent mode

The token was being obtained successfully but not outputted from the prepare step for use in later steps. Added explicit output of the GitHub token and debug logging.

.github/workflows/claude-auto-review-test.yml

@@ -18,9 +18,9 @@ jobs:
           fetch-depth: 1
 
       - name: Auto review PR
-        uses: anthropics/claude-code-action@main
+        uses: anthropics/claude-code-action@v1-dev
         with:
-          direct_prompt: |
+          prompt: |
             Please review this PR. Look at the changes and provide thoughtful feedback on:
             - Code quality and best practices
             - Potential bugs or issues
@@ -30,4 +30,4 @@ jobs:
 
             Be constructive and specific in your feedback. Give inline comments where applicable.
           anthropic_api_key: ${{ secrets.ANTHROPIC_API_KEY }}
-          allowed_tools: "mcp__github__create_pending_pull_request_review,mcp__github__add_comment_to_pending_review,mcp__github__submit_pending_pull_request_review,mcp__github__get_pull_request_diff"
+          claude_args: "--allowedTools mcp__github-comment-server__create_pending_pull_request_review,mcp__github-comment-server__add_comment_to_pending_review,mcp__github-comment-server__submit_pending_pull_request_review,mcp__github-comment-server__get_pull_request_diff"

README.md

@@ -1,6 +1,6 @@
 ![Claude Code Action responding to a comment](https://github.com/user-attachments/assets/1d60c2e9-82ed-4ee5-b749-f9e021c85f4d)
 
-# Claude Code Action
+# Claude Code Action - No Approval Test
 
 A general-purpose [Claude Code](https://claude.ai/code) action for GitHub PRs and issues that can answer questions and implement code changes. This action listens for a trigger phrase in comments and activates Claude act on the request. It supports multiple authentication methods including Anthropic direct API, Amazon Bedrock, and Google Vertex AI.
 

action.yml

@@ -141,7 +141,7 @@ runs:
         echo "Base-action dependencies installed"
         cd -
         # Install Claude Code globally
-        bun install -g @anthropic-ai/claude-code@1.0.73
+        bun install -g @anthropic-ai/claude-code@1.0.72
 
     - name: Setup Network Restrictions
       if: steps.prepare.outputs.contains_trigger == 'true' && inputs.experimental_allowed_domains != ''

base-action/action.yml

@@ -85,7 +85,7 @@ runs:
 
     - name: Install Claude Code
       shell: bash
-      run: bun install -g @anthropic-ai/claude-code@1.0.73
+      run: bun install -g @anthropic-ai/claude-code@1.0.72
 
     - name: Run Claude Code Action
       shell: bash

src/entrypoints/prepare.ts

@@ -56,6 +56,7 @@ async function run() {
       mode,
       githubToken,
     });
+    core.setOutput("GITHUB_TOKEN", githubToken);
 
     // MCP config is handled by individual modes (tag/agent) and included in their claude_args output
 

src/modes/agent/index.ts

@@ -39,9 +39,11 @@ export const agentMode: Mode = {
     return false;
   },
 
-  async prepare({ context }: ModeOptions): Promise<ModeResult> {
+  async prepare({ context, githubToken }: ModeOptions): Promise<ModeResult> {
     // Agent mode handles automation events and any event with explicit prompts
     
+    console.log(`Agent mode: githubToken provided: ${!!githubToken}, length: ${githubToken?.length || 0}`);
+
     // TODO: handle by createPrompt (similar to tag and review modes)
     // Create prompt directory
     await mkdir(`${process.env.RUNNER_TEMP}/claude-prompts`, {
@@ -72,7 +74,7 @@ export const agentMode: Mode = {
             `${process.env.GITHUB_ACTION_PATH}/src/mcp/github-comment-server.ts`,
           ],
           env: {
-            GITHUB_TOKEN: process.env.GITHUB_TOKEN || "",
+            GITHUB_TOKEN: githubToken || "",
             REPO_OWNER: context.repository.owner,
             REPO_NAME: context.repository.repo,
             GITHUB_EVENT_NAME: process.env.GITHUB_EVENT_NAME || "",

undefined/claude-prompts/claude-prompt.txt

@@ -0,0 +1 @@
+Custom prompt content