菜单
Personal observability for Claude Code — daily token/cost reports, tool usage, project stats, encrypted multi-device sync
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
| name | traceme |
| description | Personal observability for Claude Code — daily token/cost reports, tool usage, project stats, encrypted multi-device sync |
Personal Claude Code observability — track your daily token usage, cost, tool usage, and project stats.
node "${CLAUDE_PLUGIN_ROOT}/scripts/traceme-cli.mjs" report today # Daily report (e.g. --brief for compact)
node "${CLAUDE_PLUGIN_ROOT}/scripts/traceme-cli.mjs" rescan # Re-derive DB from changed transcripts (--all = full rebuild)
node "${CLAUDE_PLUGIN_ROOT}/scripts/traceme-cli.mjs" dashboard # Interactive HTML dashboard (--no-open to skip browser)
node "${CLAUDE_PLUGIN_ROOT}/scripts/traceme-cli.mjs" sync push today # Encrypt + push daily snapshot to sync repo
Other commands: stats, status, export, errors, sync <setup|pull|verify|status|forget|rebuild|set-key>.
Full CLI reference (every command, all flags, date formats like yesterday / YYYY-MM-DD / --range 7d) → reference/command-reference.md.
Multi-device sync setup & privacy → reference/sync.md. Dashboard filters → reference/dashboard.md.
Model pricing is stored at ~/.claude/traceme/model_pricing.json. On first run, a default file is created with current pricing for Claude and DeepSeek models. Edit this file to update pricing or add new models — it persists across plugin updates.
By default, report and stats show the cross-device aggregate — data merged from all your devices via encrypted sync. This is read from the locally cached origin/main git ref (no network call), so it reflects the last sync.
Cross-device data is auto-pulled at session start when sync is configured. If no cross-device data exists for the requested date (sync not set up, or nothing pushed yet), output falls back to local SQLite data. Pass --local to always force the local view.
.enc (AES-256-GCM encrypted) files — no plaintext ever touches GitHub~/.claude/traceme/key.txt and never committedDisplays today's per-project breakdown: sessions, prompt count, token usage, cost, top expensive prompts, and tool/skill usage. Use traceme export to export daily summaries as JSON or CSV.
Iterative TDD review→fix loop — each round critiques the codebase, fans out agents to fix findings, reviews the un-fixed and judges why, escalates hard calls to the human, runs the test suite before each commit (escalating on failure), and commits. Use when the user wants to drive a codebase to a clean state over multiple automated review/fix rounds.
Internal skill for presenting Codex image generation output
Internal guidance for presenting fabric MCP tool output back to the user
Post-feature sharp review (锐评) —parallel reviewers, merge findings, sync task list
REM sleep for Claude sessions — summarize what changed, update memory, crystallize if needed
Refresh stale knowledge-base docs, and create/bind new ones. Detects living docs whose doc_source subtrees have drifted (by commits, churn, or age) past their device-local anchor, then incrementally rewrites each — minimal changes, preserving human annotations — and re-anchors to HEAD. Use to bring docs up to date with code, to turn a doc into a bound doc, or after a todo report shows DOC- tasks.