Skip to main content
command CX Codex

anomalies — Claude Code command: what it does, how to install

anomalies spots outlier agent sessions with unusual token spikes or dollar costs using z-score calculations.

MIT ★ 1,198 updated 2026-10-23

Install into Claude Code · Plugin marketplace All methods and platforms ↓
/plugin marketplace add hoangsonww/Claude-Code-Agent-Monitor
/plugin install ccam-insights@claude-code-agent-monitor-plugins
Open on GitHub

What anomalies does

This command queries your local Agent Monitor dashboard at http://localhost:4820 to pull up to 200 recent sessions. It calculates the mean and standard deviation for run expenses and token usage across the retrieved population.

Any session exceeding the z-score threshold (default 2.0) is rendered into a Markdown table. Total tokens account for input, output, cache-read, and cache-write values. Items are sorted by cost deviation and flagged with 🟡 for z > 2 or 🔴 for z > 3.

If the dashboard server is down at http://localhost:4820, metric gathering fails entirely. You need to launch it with npm start from the Agent Monitor repository root before invoking this command.

What it helps with

  • spot sessions that exceed normal API cost baselines
  • identify runaway token counts across prompt loops
  • isolate cost outliers using a custom z-score threshold
  • inspect token totals including cache reads and writes

Example prompts for the agent

  • /anomalies
  • /anomalies 2.5
  • /anomalies 1.2

How to install anomalies

1Plugin marketplace

/plugin marketplace add hoangsonww/Claude-Code-Agent-Monitor
/plugin install ccam-insights@claude-code-agent-monitor-plugins

2Manually: put the files in place

.claude/commands/anomalies.md

Call it with /anomalies. For all projects use ~/.claude/commands/

Questions about anomalies

How to install anomalies in Claude Code?

Run /plugin marketplace add hoangsonww/Claude-Code-Agent-Monitor first, then run /plugin install ccam-insights@claude-code-agent-monitor-plugins.

What does the anomalies command do?

It calculates statistical z-scores over your recent agent runs and highlights executions with abnormal token or dollar expenses.

What is required before using anomalies?

An active Agent Monitor web server listening on http://localhost:4820 is required. Start it by running npm start in the repository root.

Source text

anomalies.md · MIT
---
description: List current cost and token outlier sessions via z-score
argument-hint: "[threshold]"
---

List the current cost/token **outlier** sessions from the Agent Monitor dashboard using a z-score test. **$ARGUMENTS** optionally sets the z-score threshold (default `2.0`; lower = stricter).

1. Fetch the population:
   - `curl -s "http://localhost:4820/api/sessions?limit=200"` → a session list; each item has `id`, `status`, `model`, `cwd`, `started_at`, `cost`, and `metadata`.

2. Compute the baseline over all returned sessions:
   - Mean and standard deviation of `cost`.
   - For sessions where you need token totals, pull `curl -s http://localhost:4820/api/pricing/cost/<id>` and sum `input_tokens + output_tokens + cache_read_tokens + cache_write_tokens`; compute mean and stddev of total tokens too.

3. Flag outliers: any session whose `z = (value − mean) / stddev` exceeds the threshold (default 2.0) on cost (primary) or tokens (secondary). Skip the calc gracefully if stddev is 0.

4. Print the flagged sessions, sorted by descending cost z-score:
   - Session id (short), model, started_at.
   - Cost (USD, 4 decimals) and its z-score.
   - Total tokens and its z-score (when fetched).
   - A flag tag: 🔴 if z > 3, 🟡 if z > 2.

Output rules: a Markdown table of flagged sessions only; currency in USD to 4 decimals; z-scores to 2 decimals; if nothing exceeds the threshold, say "No cost/token outliers above z=<threshold>" and report the top session by cost for context. Cite only API values — never fabricate. If the dashboard is unreachable at `http://localhost:4820`, tell the user to start it with `npm start` from the repo root.

© Son Nguyen · MIT · hoangsonww/Claude-Code-Agent-Monitor

Similar slash commands

All slash commands for Agent setup: memory, context, orchestration →

From the same set155

The whole set →

Project sponsors

Companies that keep this analytics open