cnw-stats-local-server - SKILL.md Agent Skill

name: cnw-stats-local-server description: Run a local Go HTTP server that captures Nx Cloud API traffic from create-nx-workspace (CNW) and nx init. Use when testing CNW telemetry, recordStat payloads, init analytics, or any request CNW/init would send to NX_CLOUD_API. Triggers on "test CNW stats", "test CNW analytics locally", "test nx init stats", "capture recordStat", "local nx cloud api", "log CNW requests", "test telemetry locally".

CNW / nx init Local Stats Server

Tiny Go HTTP server that intercepts every request CNW (create-nx-workspace) or nx init would send to NX_CLOUD_API. Logs method, URL, headers, and body to stdout and a log file (truncated on startup). Responds 200 {} to everything so CNW/init proceed normally.

Source: ~/projects/dot-ai-config/tools/cnw-log-server/

When to use

User wants to see what recordStat / analytics payloads CNW or nx init send.
User is testing a new flow variant, prompt copy, or telemetry code value and needs ground truth.
User says "point NX_CLOUD_API at a local server" / "log the requests CNW makes".

Run it

cd ~/projects/dot-ai-config/tools/cnw-log-server
go run . -port 8765 -log cnw-log-server.log

Flags:

-port — listen port (default 8765)
-log — log file path (default ./cnw-log-server.log, truncated on each startup)

The server prints a ready line like:

cnw-log-server listening on http://localhost:8765  log=./cnw-log-server.log
usage: NX_CLOUD_API=http://localhost:8765 <your-command>

Point CNW / init at it

In a separate terminal:

# create-nx-workspace
NX_CLOUD_API=http://localhost:8765 npx create-nx-workspace@latest myworkspace --preset=react-monorepo

# nx init (inside an existing repo)
NX_CLOUD_API=http://localhost:8765 npx nx@latest init

Every request is logged in both the terminal running the server and the log file.

What each request looks like

=== req #3  2026-04-23T09:22:13.526-04:00  POST /nx-cloud/stats  from=[::1]:56536  ct=application/json  len=120 ===
  Accept: */*
  Content-Type: application/json
  User-Agent: axios/1.6.0
  body (json):
  {
    "command": "create-nx-workspace",
    "code": "setup-nx-cloud-v2-a",
    ...
  }

Common flags for CNW/init testing

Combine with the server to exercise specific paths:

NX_CNW_FLOW_VARIANT=0|1|2 — force a flow/prompt variant (see GEMINI.md › CNW A/B Testing).
NX_CLOUD_API=http://localhost:8765 — route all stat/telemetry calls to the log server.
CI=true — simulate CI environment, skips interactive prompts.

Example — confirm the code value a variant sends:

# terminal 1
cd ~/projects/dot-ai-config/tools/cnw-log-server && go run . -port 8765

# terminal 2
NX_CNW_FLOW_VARIANT=1 NX_CLOUD_API=http://localhost:8765 \
  npx create-nx-workspace@latest /tmp/scratch --preset=react-monorepo --nxCloud=skip

Then grep '"code"' ~/projects/dot-ai-config/tools/cnw-log-server/cnw-log-server.log to confirm the variant fired the expected payload.

Notes

The server is stdlib-only; no go.sum or deps to install.
Responds 200 with {} to every path/method — CNW and init don't care about response shape beyond the status code.
Log file is truncated on each startup. If you need to keep a prior run, copy it before restarting.
Port 8765 is the default; change it if something else is bound (lsof -i :8765).