sheriff-triage

Analyze Firefox CI failures by aggregating Taskcluster data and Treeherder classifications.

3|Updated Jan 8, 2026
One-click install
npx skills add https://github.com/jwmossmoz/agent-skills --skill sheriff-triage
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: sheriff-triage
Source: https://github.com/jwmossmoz/agent-skills/tree/main/skills/sheriff-triage
Command: npx skills add https://github.com/jwmossmoz/agent-skills --skill sheriff-triage

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires treeherder-client, requests, and includes scripts (resource) components.

What problem does it solve?

Sheriff Triage automates the comprehensive failure analysis for CI tasks to quickly identify whether failures are caused by code changes, image changes, or intermittent infra issues.

Core Features & Use Cases

  • Collects Taskcluster task data (pool, status) and cross-references with Treeherder classifications to determine root causes.
  • Compares alpha/production worker images to detect image-related regressions.
  • Searches for similar failures across autoland and mozilla-central to provide contextual triage signals.
  • Produces a structured verdict with supporting rationale and recommended actions for engineers.

Quick Start

Run uv run triage.py <TASK_ID> to triage a failing task and view the report.

Frequently Asked Questions about sheriff-triage

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automate CI failure triage for a failing Taskcluster task?

Automate CI failure triage by running the triage script with a specific task ID to aggregate task data, image versions, and Treeherder classifications into a structured root-cause verdict with recommended actions.

What is the best way to find root causes for Firefox CI failures on autoland and mozilla-central?

Find root causes for Firefox CI failures by cross-referencing Taskcluster task definitions and status with Treeherder classifications, then searching for similar failures across autoland and mozilla-central branches to provide contextual triage signals.

Can I compare alpha and production worker images to detect CI image regressions?

Yes, you can compare alpha and production worker images to detect image-related regressions by reading worker SBOM and image metadata from Taskcluster during the triage process.

How does Treeherder classification integration work during continuous integration failure analysis?

Treeherder classification integration works by fetching classification data and cross-referencing it with Taskcluster task status to determine whether failures stem from code changes, image changes, or intermittent infrastructure issues.

Do I need the treeherder-client and requests dependencies to run CI failure analysis?

Yes, you need the treeherder-client and requests dependencies installed to fetch Treeherder classifications and query Taskcluster for task definitions, status, and worker image metadata during failure analysis.