Stop guessing which Claude Code model and settings to use.

Claude Code Usage Advisor analyzes your local Claude Code history and recommends practical launch profiles for Sonnet, Opus plan mode, Haiku, effort level, permission mode, sandboxing, and prompt caching.

It is read-only, zero-dependency, and runs locally against ~/.claude (the Claude Code CLI default on macOS, Linux, and Windows).

Run without installing:

npx claude-code-usage-advisor

Or install globally:

npm install -g claude-code-usage-advisor
cc-advisor

From a clone:

git clone https://github.com/LZong-tw/claude-code-usage-advisor.git
cd claude-code-usage-advisor
npm test
node bin/claude-code-usage-advisor.js

• Which model should be your default for daily Claude Code work?
• When should you use Sonnet, Opus plan mode, Haiku, or fast mode?
• Should effortLevel be global or task-specific?
• Is permissions.defaultMode appropriate for your actual usage?
• Are risky allow rules accumulating in settings.json?
• Should you configure autoMode.environment, sandboxing, prompt caching, or subprocess env scrubbing?
• Which launch profiles should you use for planning, implementation, autonomous work, and cheap triage?
• Where are tool errors, hooks, status lines, project hotspots, MCP calls, or global context creating friction?

First run, full report against your real ~/.claude:

npx claude-code-usage-advisor

That's the answer for most people. The sections below cover the rest.

Audit only your recent shift in workflow. Default window is 30 days; narrow when you've changed how you use Claude Code lately and want recommendations based on that, not on history.

cc-advisor --days 7

Share the report with a teammate or paste into a ticket. Produces a self-contained HTML file (no external assets) you can open in a browser or upload as an artifact.

cc-advisor --html reports/claude-code-insights.html

Feed it into a dashboard or pipeline. JSON contains the same recommendations, evidence, and investigations as the text report.

cc-advisor --json > advisor.json
jq '.recommendations[] | select(.severity=="high")' advisor.json

Speed up runs on a huge ~/.claude. Scans newest JSONL files first.

cc-advisor --days 14 --max-files 200

Point at a non-default Claude Code directory — for example a teammate's exported ~/.claude snapshot, or a backup directory you're auditing offline:

cc-advisor --claude-dir ./teammate-claude-dump
cc-advisor --claude-dir /mnt/backup/2026-05/claude

Skip the settings snippets when you just want the diagnosis, not the suggested JSON:

cc-advisor --no-snippets

--claude-dir <path>   Claude Code directory (default: ~/.claude)
--days <n>            Analyze JSONL records from the last n days (default: 30, 0 = all)
--max-files <n>       Cap scanned JSONL files, newest first (default: unlimited)
--json                Print machine-readable JSON
--html <path>         Write a self-contained HTML report
--no-snippets         Hide settings snippets in text output
--help                Show help

Abridged from a real run. The full report also includes Current settings, Lifetime model usage, Recent usage patterns, Top tools / Bash command families, and (unless --no-snippets) ready-to-paste settings.json blocks.

Additional investigations
-------------------------
[high] workflow-friction: Investigate tool error hot spots
  evidence: Tool error rate is 33.3% (1/3). Bash 50.0% (1/2)
  action: Look at failed Bash/Edit/WebFetch patterns first. Repeated tool errors
    usually mean missing project scripts, stale permissions, brittle hooks, or
    prompts that ask Claude to guess commands instead of inspecting repo affordances.

Recommended launch profiles
---------------------------
- Daily implementation: `claude --model sonnet --permission-mode acceptEdits --effort medium`
- Deep planning then execution: `claude --model opus --permission-mode plan --effort xhigh`
- Trusted autonomous work: `claude --model sonnet --permission-mode auto --effort high`
- Large-context planning: `claude --model opus --permission-mode plan --effort xhigh --add-dir <extra-dir>`
- Cheap/simple triage: `claude --model haiku --permission-mode default --effort low`

Recommendations
---------------
[high] model: Make Sonnet the execution default and reserve Opus for planning
  evidence: Opus is 99.7% of non-cache tokens while recent sessions are
    tool-heavy (1.50 tool calls/assistant call).
  action:   Use `claude --model opus --permission-mode plan --effort xhigh` for
    ambiguous work, then let execution run on Sonnet.

[high] permissions: Move risky always-allow permissions to ask/deny
  evidence: Detected 2 critical/high-risk allow rules, including destructive
    shell, cloud/cluster mutation, secret, or network pipe-to-shell patterns.
  action:   Keep read-only commands in `allow`; move deploy, destructive, secret,
    and broad network shell commands to `ask`.

Every recommendation cites the evidence it's based on, so you can decide whether a heuristic fits your situation before applying the suggested action.

The tool reads local files only:

• ~/.claude/settings.json
• ~/.claude/stats-cache.json
• ~/.claude/projects/**/*.jsonl

It does not call Anthropic APIs, GitHub APIs, or any network service. It does not modify your settings.

The advisor combines:

• Lifetime model token totals from stats-cache.json
• Recent model/tool/permission mode patterns from JSONL transcripts
• Risk classification for current permissions.allow rules
• Heuristics from Claude Code's current model and settings behavior

Core routing policy:

• sonnet: daily implementation, tests, straightforward refactors
• opus with --permission-mode plan: ambiguous planning, architecture, migrations, high-stakes debugging
• Opus large-context variants: large-context planning when your account and Claude Code CLI expose a supported full model id
• haiku: summaries, classification, cheap triage
• fast mode: explicit, urgent Opus sessions only, not a global default

This tool is designed for public use on private machines:

• No telemetry
• No dependency install scripts
• No external network calls
• No settings writes
• No transcript content is printed by default, only aggregate counts and command families

The JSON output may include project paths, command families, model names, tool names, and settings keys. Review it before sharing publicly.

npm test
npm run smoke

No build step is required.

MIT