aws / aws/amazon-q-developer-cli

MCP servers not loading due to "Agent missing path info" error

Open
#2,452 20 comments 4 reactions 0 assignees View on GitHub
Dominant language
Rust
Stars
2k
Forks
439
PR merge metrics
No merged PRs in 30d

Description

### Checks

- [x] I have searched [github.com/aws/amazon-q-developer-cli/issues](https://github.com/aws/amazon-q-developer-cli/issues?q=) and there are no duplicates of my issue
- [x] I have run `q doctor` in the affected terminal session
- [x] I have run `q restart` and replicated the issue again

### Operating system

macOS 15.5.0 (24F74)

### Expected behaviour

MCP servers should load properly and be available for use in Q chat. The `q mcp list` command should show configured servers with their status.

### Actual behaviour

MCP servers fail to load with "Agent missing path info" error. When running `q mcp list`, it returns "error: Agent missing path info". The `q agent list` command shows "q_cli_default **No path found**" indicating the default agent configuration is corrupted or missing path information.

```
[chat-transcript]
Great! I can see that `export PATH="$HOME/.local/bin:$PATH"` is already in your `.zshrc` file. However, the Q CLI agent migh

The issue is that we're running bash but trying to source a zsh configuration file. Let's check the current PATH to see if `~/.local/bin` is there:
[Tool uses: execute_bash]

The PATH issue is still there. Let's manually add `~/.local/bin` to the PATH for this session and then restart the Q CLI:
[Tool uses: execute_bash]

[Tool uses: execute_bash]

[Tool uses: execute_bash]

The PATH export didn't work as expected. Let me try a different approach. The issue might be that the Q CLI agent is running with a different environment. Let's try to kill all Q CLI processes and restart them:
[Tool uses: execute_bash]

[Tool uses: execute_bash]

The processes are still running. Let me try a more forceful approach:
[Tool uses: execute_bash]

Now let's try a different approach. The issue might be that the MCP servers configuration is pointing to the wrong path for `uvx`. Let's modify the MCP configuration to use the full path:
[Tool uses: fs_read]

> you've tried all of this before. i dont understand how your edits arent working

You're absolutely right - I apologize for going in circles. Let me take a step back and think about this differently.

The core issue is that the Q CLI is giving an "Agent missing path info" error, which suggests there's a fundamental problem with how the Q CLI agent is configured or running, not just a PATH issue.

Let me try a few more targeted approaches:

1. **Check if there's an agent configuration issue:**
[Tool uses: execute_bash]

2. **Try updating Q CLI to see if it's a version issue:**
[Tool uses: execute_bash]

[Tool uses: execute_bash]

3. **The "Agent missing path info" error might be a known issue. Let me check if there's a way to reinitialize the agent:**
[Tool uses: execute_bash]

This is interesting! There are agent commands. Let me check what agents are available:
[Tool uses: execute_bash]

There it is! The agent `q_cli_default` has "**No path found**" - this is exactly the "Agent missing path info" issue. This suggests the default agent configuration is broken.

Let me try to see if we can fix this by setting a default agent or recreating the agent configuration:
[Tool uses: execute_bash]

The issue seems to be that the default agent configuration is missing its path. This might be a bug or corruption in the Q CLI installation.

At this point, I think the best approach would be to:

1. **Report this as a bug** since "Agent missing path info" with "No path found" suggests a corrupted installation
2. **Try reinstalling Q CLI** to fix the agent configuration

Would you like me to help you report this issue or try a reinstallation? The MCP servers themselves seem to be configured correctly - the problem is with the Q CLI agent system that manages them.
[Tool uses: none]

> both please

I'll help you with both - reporting the issue and trying a reinstallation.

## 1. First, let's report this as a bug:
[Tool uses: report_issue]

(...truncated)
```

### Steps to reproduce

1. Configure MCP servers in ~/.config/q/mcp-servers.json (servers are properly configured with valid paths)
2. Run `q mcp list` - returns "Agent missing path info" error
3. Run `q agent list` - shows "q_cli_default **No path found**"
4. Run `q doctor` - shows PATH issues but MCP servers still don't load even after fixing PATH
5. MCP server executables exist and are accessible (uvx, builder-mcp, amzn-mcp all found in expected locations)

### Environment

```yaml

[q-details]
version = "1.13.0"
hash = "7dee6efdf68f6343e80a56028e6b40f8eff77746"
date = "2025-07-31T02:57:26.581678Z (21h ago)"

[system-info]
os = "macOS 15.5.0 (24F74)"
chip = "Apple M1 Pro"
total-cores = 10
memory = "32.00 GB"

[environment]
cwd = "/Users/USER/10_ATOM_Active/current"
cli-path = "/Users/USER/10_ATOM_Active/current"
install-method = "toolbox"

[env-vars]
PATH = "/opt/homebrew/bin:/opt/homebrew/sbin:/Users/USER/.toolbox/bin:/usr/local/bin:/System/Cryptexes/App/usr/bin:/usr/bin:/bin:/usr/sbin:/sbin:/var/run/com.apple.security.cryptexd/codex.system/bootstrap/usr/local/bin:/var/run/com.apple.security.cryptexd/codex.system/bootstrap/usr/bin:/var/run/com.apple.security.cryptexd/codex.system/bootstrap/usr/appleinternal/bin:/usr/local/go/bin:/Users/USER/.cargo/bin:/Applications/Visual Studio Code.app/Contents/Resources/app/bin:/opt/homebrew/opt/fzf/bin:/Users/USER/Library/Application Support/Code/User/globalStorage/github.copilot-chat/debugCommand"
SHELL = "/bin/zsh"
TERM = "xterm-256color"
ZDOTDIR = "/Users/USER"
__CFBundleIdentifier = "com.microsoft.VSCode"

[chat-settings]

[chat-trusted_tools]
execute_bash=trusted
fs_read=trusted

[chat-failed_request_ids]
none

[chat-context]
current_profile=q_cli_default
profile_context=
AmazonQ.md
README.md
.amazonq/rules/**/*.md

files=none
```

Contributor guide

Open the contributing guide

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.