metric-rca

metric-rca is a command for coding agents from mixpanel/ai-plugins. It costs 0 tokens per session (4,738 once invoked), scanned A, original, Apache-2.0.

A command that investigates why a metric changed after an earlier anomaly or drift diagnosis. It splits the data into different groups to find where the change was concentrated.

In plain words
What is it for?
Use it after metric-anomaly or metric-drift has produced a diagnosis. It ranks the findings and adds them to an existing Mixpanel diagnosis board.
Why use it?
It helps explain the cause of an already-confirmed metric movement instead of making you inspect every segment manually. It does not detect anomalies itself.

Command

Part of the mixpanel plugin — 12 skills, 23 commands shipped together

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add commands/mixpanel/ai-plugins/metric-rca
Clone the repo
git clone --depth 1 https://github.com/mixpanel/ai-plugins

Or install mixpanel, the plugin that ships this one along with the rest of its 12 skills, 23 commands.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for metric-rca

README.md
[![agentmods](https://agentmods.dev/badge/commands/mixpanel/ai-plugins/metric-rca.svg)](https://agentmods.dev/commands/mixpanel/ai-plugins/metric-rca)
Your own site
<a href="https://agentmods.dev/commands/mixpanel/ai-plugins/metric-rca"><img src="https://agentmods.dev/badge/commands/mixpanel/ai-plugins/metric-rca.svg" alt="Measured on agentmods" height="20"></a>
Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 4,738 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00000 $0.04738
Opus 5 $0.00000 $0.02369
Sonnet 5 $0.00000 $0.00948
Haiku 4.5 $0.00000 $0.00474

Measured 5d ago against content hash 3e5339c5ab91, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

metric-rca scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 5d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

plugins/mixpanel/skills/monitor-metrics/commands/metric-rca.md · 341 lines

How it starts

The opening of the file, as written. The whole thing — 341 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Command: metric-rca

Root-cause investigation for a flagged metric. Takes the diagnosis payload from a prior metric-anomaly or metric-drift run and fans out across a set of segmentation branches to localise where the movement concentrated. Produces a ranked list of findings and appends them to the diagnosis board the user already created.

This command does not re-run anomaly or drift detection. It assumes the movement has already been established — its job is attribution, not detection.


Prerequisites

Before this command runs, the session must hold a diagnosis payload in conversation memory from an earlier metric-anomaly or metric-drift run (see references/execution.md Step 2). The payload carries the project, metric, metric type, date ranges, flagged points or drift windows, and the query bodies used.

If no payload exists, do not attempt to run RCA from a cold start. Tell the user: "RCA runs on top of an existing anomaly or drift diagnosis. Run metric-anomaly or metric-drift first, then come back here." Stop.

Board state

If the user persisted the diagnosis as a Mixpanel board (Step 2 in references/execution.md), the payload will include diagnosis_board_id. This command appends to that board — it does not create a new one. If no board was created, skip the append step at the end and just return the findings inline; do not silently create a new board.

Ask once — business / market context

Before firing Branch 5, ask the user exactly once:

"What business or market is this metric tied to? (e.g., Indian e-commerce, Indian OTT streaming, SEA fintech.) I'll use this to check whether the flagged dates line up with festivals, launches, or category-specific events."

Hold the answer as business_context. If the user skips or says "not relevant", skip Branch 5 entirely — do not guess the market from project name or memory.


Phase 1 — Branch selection + parallel fan-out

Read the payload and decide which branches to run. Every branch runs against the same date ranges the source command used:

Read the full file on GitHub · 341 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 5d ago First seen · 341 lines · 0 tokens per session scan A 3e5339c5ab91

Subscribe to this mod's changes

metric-rca is a command published in the GitHub repository mixpanel/ai-plugins (15 stars, last pushed yesterday), licensed Apache-2.0. It costs nothing until one of its globs matches a file; then it loads 4,738 tokens. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.