Not able to create session with rovodev

[Amogh Agrawal]

I was able to use acli rovodev run for last month and half, but It has stopped working since last update. I am using latest 1.3.1-stable.

I did chmod 755/777 for executable, still the same. I reinstalled using brew on mac. still the same. I have deleted my .rovodev folder to get it started fresh, still nothing happens.

Here are logs:

2025-08-22 18:10:23.643 | INFO | - Starting Rovo Dev CLI
2025-08-22 18:10:23.643 | INFO | - Working directory: /Users/<user>/ws
2025-08-22 18:10:23.643 | INFO | - Config file: /Users/<user>/.rovodev/config.yml
2025-08-22 18:10:24.039 | WARNING | - Failed to retrieve dynamic configuration, using defaults.
2025-08-22 18:10:24.039 | WARNING | - Failed to retrieve dynamic configuration, using defaults.
2025-08-22 18:10:24.039 | INFO | - Starting new session
2025-08-22 18:10:24.278 | WARNING | - Failed to retrieve dynamic configuration, using defaults.
2025-08-22 18:10:24.278 | WARNING | - Failed to retrieve dynamic configuration, using defaults.
2025-08-22 18:10:24.302 | WARNING | - Failed to retrieve dynamic configuration, using defaults.
2025-08-22 18:10:24.302 | WARNING | - Failed to get model for ID: claude-sonnet-4@20250514, falling back to default model configuration.
2025-08-22 18:10:24.302 | WARNING | - Failed to retrieve dynamic configuration, using defaults.
2025-08-22 18:10:24.324 | WARNING | - Failed to retrieve dynamic configuration, using defaults.
2025-08-22 18:10:24.324 | INFO | - Starting MCP servers
2025-08-22 18:10:24.328 | WARNING | - Failed to retrieve dynamic configuration, using defaults.
2025-08-22 18:10:56.565 | ERROR | - Failed to start MCP server {'command': '/Users/<user>/.local/share/acli/1.3.1-stable/plugin/rovodev/atlassian_cli_rovodev', 'args': ['nautilus', 'run', '--tools', 'open_files,create_file,delete_file,move_file,expand_code_chunks,find_and_replace_code,grep,bash,powershell,update_allowed_external_paths', '--workspace-args-json', '{"workspace_view_max_files": 1000, "allowed_external_paths": [], "run_shell_in_sandbox": false}'], 'env': None, 'cwd': None, 'tool_prefix': None, 'log_level': None, 'log_handler': None, 'timeout': 30, 'process_tool_call': None, 'allow_sampling': True, 'exclude_tools': [], 'config_path': None, 'name': 'code-nautilus', 'log_file': None, 'transport': 'stdio'}
2025-08-22 18:10:56.565 | ERROR | - Failed to start built-in MCP server - Failed to start STDIO MCP server with command '/Users/<user>/.local/share/acli/1.3.1-stable/plugin/rovodev/at...

If you continue to see this error, please report it to the Rovo Dev team: [blue underline]https://rovodevagents.atlassian.net/servicedesk/customer/portal/1/group/1/create/45[/blue underline]
2025-08-22 18:10:56.567 | INFO | - To resume your session, restart Rovo Dev CLI with the --restore flag.




Amazon Q produced this report after trying to fix it:

## **BUG REPORT: acli rovodev MCP Server Fails Due to SQLite Database Errors**

### **System Information:**

Darwin MAC-F42HM4FPPV 24.6.0 Darwin Kernel Version 24.6.0: Mon Jul 14 11:29:54 PDT 2025; root:xnu-11417.140.69~1/RELEASE_ARM64_T8122 arm64

acli version: 1.3.3-stable

Rovo Dev CLI: 0.11.16

Operating System: Darwin 24.6.0 (arm64)

### **Issue:**

MCP server consistently fails to start with 30-second timeout. Agent starts, shows input box, then crashes with "Failed to start STDIO MCP server" error.

### **Root Cause:**

Massive SQLite database errors in atlassian_cli_rovodev binary:

### **SQLite Error Log Sample:**

atlassian_cli_rovodev: (libsqlite3.dylib) API call with invalid database connection pointer

atlassian_cli_rovodev: (libsqlite3.dylib) misuse at line 186167 of [1b37c146ee]

atlassian_cli_rovodev: (libsqlite3.dylib) close deferred due to unfinalized statement: "PRAGMA synchronous"

### **Rovodev Log:**

2025-09-05 20:20:28.459 | ERROR | Failed to start MCP server {'command': '/Users/amoghtest2/.local/share/acli/1.3.3-stable/plugin/rovodev/atlassian_cli_rovodev', 'args': ['nautilus', 'run', '--tools', 'open_files,create_file,delete_file,move_file,expand_code_chunks,find_and_replace_code,grep,bash,powershell,update_allowed_external_paths', '--workspace-args-json', '{"workspace_view_max_files": 1000, "allowed_external_paths": [], "run_shell_in_sandbox": false}'], 'timeout': 30}

### **Reproduction:**

1. Fresh install: brew uninstall atlassian/acli/acli && brew install atlassian/acli/acli

2. Clean setup: rm -rf ~/.rovodev ~/.config/acli ~/.local/share/acli

3. Authenticate: acli rovodev auth login

4. Run: acli rovodev run

5. Result: Agent starts then crashes after 30 seconds

### **Additional Notes:**

• Issue persists across complete reinstalls

• SQLite errors occur immediately when MCP server starts

• Problem exists in the binary itself, not user configuration

• System Integrity Protection is enabled

This is a critical software defect in acli rovodev 1.3.3-stable affecting macOS arm64 systems.

[David Lagacé]

I have the same error. In the log files, I've find out : 

.........

DATE | WARNING  | - Request error: [SSL: CERTIFICATE_VERIFY_FAILED] certificate verify failed: self-signed certificate in certificate chain (_ssl.c:1028). Retrying...

........

Seem related to a proxy connectivity error. Are you behind a corpo proxy ?

[Amogh Agrawal]

I am behind a proxy, but it has been until now. I asked Amazon q to investigate and here is the bug report produced by Q:

## **BUG REPORT: acli rovodev MCP Server Fails Due to SQLite Database Errors**

### **System Information:**

Darwin MAC-F42HM4FPPV 24.6.0 Darwin Kernel Version 24.6.0: Mon Jul 14 11:29:54 PDT 2025; root:xnu-11417.140.69~1/RELEASE_ARM64_T8122 arm64

acli version: 1.3.3-stable

Rovo Dev CLI: 0.11.16

Operating System: Darwin 24.6.0 (arm64)

### **Issue:**

MCP server consistently fails to start with 30-second timeout. Agent starts, shows input box, then crashes with "Failed to start STDIO MCP server" error.

### **Root Cause:**

Massive SQLite database errors in atlassian_cli_rovodev binary:

### **SQLite Error Log Sample:**

atlassian_cli_rovodev: (libsqlite3.dylib) API call with invalid database connection pointer

atlassian_cli_rovodev: (libsqlite3.dylib) misuse at line 186167 of [1b37c146ee]

atlassian_cli_rovodev: (libsqlite3.dylib) close deferred due to unfinalized statement: "PRAGMA synchronous"

### **Rovodev Log:**

2025-09-05 20:20:28.459 | ERROR | Failed to start MCP server {'command': '/Users/amoghtest2/.local/share/acli/1.3.3-stable/plugin/rovodev/atlassian_cli_rovodev', 'args': ['nautilus', 'run', '--tools', 'open_files,create_file,delete_file,move_file,expand_code_chunks,find_and_replace_code,grep,bash,powershell,update_allowed_external_paths', '--workspace-args-json', '{"workspace_view_max_files": 1000, "allowed_external_paths": [], "run_shell_in_sandbox": false}'], 'timeout': 30}

### **Reproduction:**

1. Fresh install: brew uninstall atlassian/acli/acli && brew install atlassian/acli/acli

2. Clean setup: rm -rf ~/.rovodev ~/.config/acli ~/.local/share/acli

3. Authenticate: acli rovodev auth login

4. Run: acli rovodev run

5. Result: Agent starts then crashes after 30 seconds

### **Additional Notes:**

• Issue persists across complete reinstalls

• SQLite errors occur immediately when MCP server starts

• Problem exists in the binary itself, not user configuration

• System Integrity Protection is enabled

This is a critical software defect in acli rovodev 1.3.3-stable affecting macOS arm64 systems.