Claude Code session retrospective: real-time waste interception, cost analytics, tool call browser, Magic Create script generator, and actionable tips.
Run all analyzers and output combined JSON results.
Extract and analyze tool chain patterns from sessions
Compare this week vs last week: cost, sessions, frustrations.
Show current config values — pricing, thresholds, hints, data paths. Verify overrides from config.env are loading.
Cost breakdown by project, model, and time period with what-if scenarios.
Matches all tools
Hooks run on every tool call, not just specific ones
Executes bash commands
Hook triggers when Bash tool is used
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.
Know what your Claude Code sessions actually cost.
Claude Code doesn't show what you're spending. No cost dashboard, no warning at 300 tool calls, no signal you used Opus for a task Sonnet could handle. cc-retrospect fixes that.
v3.0.0-rc4 — Agent Diary calendar view, tool call history browser, Magic Create script generator, STYLE.md live sync, chain pattern analysis, 5 themes, command palette.
In Claude Code:
/marketplace add vaddisrinivas/cc-retrospect
/install cc-retrospect@cc-retrospect
That's it. Hooks auto-register. Dashboard starts on first /cc-retrospect:dashboard.
git clone https://github.com/vaddisrinivas/cc-retrospect ~/.claude/plugins/cc-retrospect
cd ~/.claude/plugins/cc-retrospect && pip install -e .
/install cc-retrospect@cc-retrospect
Or: cd ~/.claude/plugins/cc-retrospect && git pull && pip install -e .
Live dashboard at localhost:7731 — open with /cc-retrospect:dashboard.
| Feature | Details |
|---|---|
| Budget tracking | Real-time spend gauge, 3 alert tiers (warning/critical/severe), 7-day sparkline, macOS notifications |
| AI savings tips | Actionable tips with $/mo projections — model switching, session length, agent replacement |
| Session grading | Every session scored A–D on efficiency, cost velocity, cache rate, waste patterns |
| Grade streak | Consecutive A-grade counter — gamifies good habits |
| Daily cost chart | Stacked bar by model (Opus/Sonnet/Haiku) per day |
| Model split donut | Visual breakdown of spend by model |
| Cost by project | Horizontal bars ranked by spend |
| Week-over-week badges | Cost, sessions, frustrations, avg duration — % change from last week |
| Session table | Searchable, sortable, filterable by project/grade/model/date |
| Session expansion | Click any row to see full session details inline |
| Agent Diary | Calendar view of daily sessions with generated notes, top projects, top tools, session trail, and spend intensity |
| Activity heatmap | Sessions by hour × day-of-week, click to filter |
| Tool usage chart | Bar chart of tool calls, click any tool to filter sessions |
| Frustration tracking | Word cloud of frustration signals (no, again, ugh, wrong, sigh) |
| Tool call browser | Browse every tool call across sessions — filter by tool, errors only |
| Chain pattern analysis | Most-used tool chains ranked by frequency and depth |
| Period comparison | Any two date ranges side-by-side with % change on all metrics |
| Profile card | Developer archetype, trait scores, fun facts — export as PNG, animated GIF, or clipboard |
| AI insights | Claude-powered analysis of session patterns (on demand) |
| Saved reports | Timestamped snapshots with open/load buttons |
| Intervention badges | Repetitive chains, oversized prompts flagged with WASTE badges |
| Config editor | Edit config.env directly from the dashboard |
| Feature | Details |
|---|---|
| 5 themes | Dark, Light, Nord, Solarized, Cyberpunk — press d to cycle |
| Command palette | ⌘K / Ctrl+K — searchable command list with 17 actions |
| Keyboard shortcuts | 1-6 jump sections, / search, r refresh, ↑↓ navigate, Enter expand |
| Glass morphism | Backdrop blur, ambient orbs, fractal noise overlay |
| GSAP animations | ScrollTrigger entrance animations on all sections |
| Responsive | Breakpoints at 1100px and 700px |
| Sidebar collapse | Toggle with hamburger menu |
| Hook | What it does |
|---|---|
| Session end | Cache cost/tokens/tools, track daily spend, log waste flags, update trends, auto-sync STYLE.md |
| Session start | Show last-session recap, daily digest (once/day), tips if thresholds exceeded |
| Pre-tool | Warn on GitHub WebFetch (use gh), Agent for simple searches (use Grep) |
| Post-tool | Dual-tier compact nudge (150+/300+ calls), auto-compact option, model nudge for simple tasks |
| User prompt | Detect mega-pastes (>1000 chars) and oversized prompts |
| Pre-compact | Log compaction with token counts |
| Post-compact | Record compaction results |
npx claudepluginhub vaddisrinivas/cc-retrospect --plugin cc-retrospecthtop for your Claude Code sessions — real-time cost tracking, cache efficiency, model comparison, smart alerts, web dashboard, and session analytics
Analyze Claude Code session token usage, costs, trends, and anomalies
Tracks Claude Code session usage (tokens, costs, skills, MCP, subagents) and sends data to a self-hosted dashboard.
Zero-setup cost tracking and budget management for Claude Code. Real-time burn rate, budget guardrails, ASCII reports, and pre-task cost estimation — all without external dependencies.
Real-time session dashboard for Claude Code — cost tracking, analytics, and smart alerts
Session analytics hook for Claude Code — 15 configurable sections (models, cost, cache, tools, files, git diff, RTK savings...)