By Snowfly0709
发呆 — track how long Claude makes you stare blankly at the screen. Logs per-turn wall time and splits it into Claude-side waiting, tool execution, and time you spent answering its questions.
发呆 — how long does Claude Code actually make you wait?
A Claude Code plugin that times every turn and splits the duration into three buckets:
AskUserQuestion, ExitPlanMode (this is you, not Claude)Logs every session, surfaces stats on demand. No telemetry, no network — everything stays in ~/.claude/.
/plugin marketplace add Snowfly0709/Claude_MindGap
/plugin install mindgap@mindgap
That's it. Hooks register automatically. The next prompt you send starts being measured.
Ask Claude naturally:
/mindgap:statsClaude will run the stats script and present a table:
Log: /Users/you/.claude/mindgap.jsonl
Sessions: 3 Turns: 47 Showing: 20
started session turn wait auto userQ tools
--------------------------------------------------------------------------------
05-18 14:02:11 a3b1c0d4 1m12s 48.30s 23.70s 0ms Bash×3(23.7s)
05-18 14:05:33 a3b1c0d4 2m04s 1m08s 55.40s 0ms Bash×4(40.2s), Read×2(15.2s)
…
Aggregate over ALL completed turns in scope:
turn duration mean= 1m24s median= 1m02s max= 4m11s total= 49m22s
Claude wait mean=42.13s median=38.20s max= 2m08s total=33m20s
auto tools mean=22.40s median=18.50s max= 1m44s total=17m32s
user-input tools mean= 2.15s median= 0ms max=39.56s total= 1m41s
--all — every turn, not just last 20--since 24h / --since 7d — time window--session <prefix> — filter by session id (8-char prefix shown in table)--include-archive — also include last week's archived log--raw — JSON output for further mathFor each hook fire, one JSONL row in ~/.claude/mindgap.jsonl:
{"ts": 1715000000.123, "event": "PreToolUse", "session_id": "abc-…", "cwd": "…", "tool": "Bash"}
Events captured: UserPromptSubmit, PreToolUse, PostToolUse, Stop. Nothing else — no prompt text, no tool inputs, no responses.
On the first hook fire of each ISO week (local time), the previous week's log moves to ~/.claude/mindgap.jsonl.prev and a fresh log starts. Anything older than ~2 weeks is gone. Read both files with --include-archive.
/plugin uninstall mindgap@mindgap
To also wipe data:
rm -rf ~/.claude/mindgap.jsonl ~/.claude/mindgap.jsonl.prev ~/.claude/mindgap/
Node.js (ships with Claude Code). No other dependencies.
MIT. See LICENSE.
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
npx claudepluginhub snowfly0709/claude_mindgap --plugin mindgapUltra-compressed communication mode. Cuts ~75% of tokens while keeping full technical accuracy by speaking like a caveman.
Memory compression system for Claude Code - persist context across sessions
Multi-model consensus engine integrating OpenAI Codex CLI, Gemini CLI, and Claude CLI for collaborative code review and problem-solving.
Curate auto-memory, promote learnings to CLAUDE.md and rules, extract proven patterns into reusable skills.