Diagnostics
Diagnostics and Issue Reports
Run the right checks before asking for help, separate daemon problems from client problems, and share only redacted output.
At a glance
01
Start diagnostics with /wenlan, wenlan status, wenlan doctor, and an MCP dry-run for the client that fails.
02
Good issue reports include environment and redacted diagnostics, not private memory contents or full ~/.wenlan archives.
01
Run the short checklist
In Claude Code, run /wenlan for connection status. In a macOS or Linux terminal, record the runtime version, status, doctor output, and a dry run for the affected client. The example uses Codex; substitute the supported client you use. The port check identifies a listener; do not stop it until you know what it is.
~/.wenlan/bin/wenlan --version
~/.wenlan/bin/wenlan status
~/.wenlan/bin/wenlan doctor
~/.wenlan/bin/wenlan connect codex --dry-run
lsof -nP -iTCP:7878 -sTCP:LISTEN02
Read the failure layer
Record whether the daemon check fails or the daemon is healthy but the client lacks tools. This distinction helps someone reproduce the problem; follow the troubleshooting guide for safe fixes.
03
Port 7878 conflicts
If the report involves port 7878, record which process is listening and whether it is a Wenlan daemon. Do not stop an unidentified process.
04
What to include in an issue
Include your Wenlan version, affected AI tool name and version, operating system, installation method, exact steps to reproduce, expected and actual result, and redacted diagnostics. Note whether the problem occurs in another AI tool.
05
What to redact
Do not paste private captures, customer or organization names, repo names, API keys, tokens, full session logs, or full ~/.wenlan archives into public issues.
Replace sensitive project text with placeholders while keeping the structure of the failure visible. If the report is security-sensitive, use the private advisory path instead.
Next
FAQ
Short answers to the adoption questions people ask before and after installing Wenlan.
Read next