flog-inspect

star 4

Use when a user asks an agent to inspect Flutter app logs, network traffic, SSE/WebSocket streams, current page state, screenshots, or debugging context through flog instead of copying from the TUI.

shaominngqing By shaominngqing schedule Updated 6/5/2026

name: flog-inspect description: Use when a user asks an agent to inspect Flutter app logs, network traffic, SSE/WebSocket streams, current page state, screenshots, or debugging context through flog instead of copying from the TUI.

flog Inspect

Use the flog ai CLI. Do not reimplement the flog wire protocol.

When Triggered

Users may not say "use flog" or name this skill. If they ask you to inspect a Flutter app's current state, logs, HTTP requests, SSE/WebSocket traffic, loading issue, button/page behavior, or screenshot-worthy UI problem, use this workflow proactively.

If device/app selection is ambiguous, first run the smallest snapshot without guessing. If it reports multiple apps, rerun with --device and/or --app from the error's next actions.

Workflow

  1. Decide if the request is visual. If the user mentions page, screen, UI, loading, button, layout, or current state, include --screenshot.
  2. Start small: run flog ai snapshot --format json --last 20 --net-last 20. Add --screenshot for visual requests.
  3. If ok=false, explain error.code, message, and next_actions.
  4. If ok=true, inspect summary, notable, logs, network, and screenshot.
  5. If the snapshot is not enough, narrow first:
    • Logs: flog ai logs --level error --last 20, flog ai logs --tag Auth --last 20, or flog ai logs --search timeout --last 20.
    • Network: flog ai net --failed --last 20, flog ai net --status 5xx --last 20, flog ai net --method POST --url login --last 20, or flog ai net --slow 1000ms --last 20.
  6. Use flog ai get <id> --detail only after choosing a stable id from snapshot/list output.
  7. Use flog ai curl net#42 when the user asks how to reproduce a request.
  8. Do not use --no-redact unless the user explicitly approves exposing secrets.
  9. Cite stable ids such as log#188, net#42, and chunk#42.13.
  10. Separate visual observations from log/network conclusions.

Response Pattern

  • Lead with what the app is doing now, not with the command output.
  • Cite the evidence ids you used.
  • Say when data is incomplete, for example no connected app, no failed requests, replay buffer missing an id, or screenshot capture failed.
  • Ask the user to reproduce the action only when the current snapshot/list/detail does not contain enough evidence.

Commands

flog ai snapshot --format json --last 20 --net-last 20
flog ai snapshot --format json --last 20 --net-last 20 --screenshot
flog ai logs --level error --last 20
flog ai net --failed --last 20
flog ai get net#42 --detail
flog ai curl net#42
flog ai watch --duration 30s --format ndjson
flog ai doctor --format json
Install via CLI
npx skills add https://github.com/shaominngqing/flog --skill flog-inspect
Repository Details
star Stars 4
call_split Forks 1
navigation Branch main
article Path SKILL.md
More from Creator
shaominngqing
shaominngqing Explore all skills →