debugging-ci-failures

Diagnoses failing GitHub Actions CI runs for PostHog pull requests and branches.

713|118|Updated Aug 11, 2020
One-click install
npx skills add https://github.com/PostHog/posthog-foss --skill debugging-ci-failures
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: debugging-ci-failures
Source: https://github.com/PostHog/posthog-foss/tree/main/.agents/skills/debugging-ci-failures
Command: npx skills add https://github.com/PostHog/posthog-foss --skill debugging-ci-failures

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Figuring out why CI is red on a PostHog pull request or on master requires digging through GitHub Actions logs, merge-queue branches, and flaky-test history by hand. This Skill triages failing runs, classifies the failure, and reports the smallest actionable finding without rerunning CI or posting to GitHub.

Core Features & Use Cases

  • CI insights digest: Starts from hogli ci:insights to see cross-run failure history and classify failures as breaking master, blocking the merge queue, flaky, or PR-only.
  • Failure classification: Maps log signals to classes such as code regression, flaky test, lint, typecheck, snapshot diff, migration, codegen drift, and infra/runner issues.
  • Safe read-only inspection: Uses gh and engineering-analytics MCP tools for log analysis while forbidding reruns, comments, pushes, or workflow edits without approval.
  • Use Case: A developer asks why their PR was kicked from the merge queue; the Skill locates the failing trunk-merge run, extracts the failing job and step, classifies it, and suggests the narrowest local reproduction command.

Quick Start

Ask the assistant to investigate why CI is failing on your current PostHog pull request and classify the failure.

Frequently Asked Questions about debugging-ci-failures

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

FAQPage Schema
How do I debug a failing GitHub Actions CI run on a PostHog PR?

Start with the hogli ci:insights digest to see cross-run failure history, then use gh run view with --log-failed to inspect the specific failing job. Classify the failure using the signal table and reproduce it locally with the narrowest hogli command.

Why was my PR kicked from the merge queue when its own checks are green?

The merge queue tests your PR on a trunk-merge branch combined with other queued PRs, running full test matrices your PR's own CI may have skipped. Inspect the failing run on the trunk-merge branch via the Trunk Merge Queue check, not gh pr checks.

How do I tell if a CI failure is flaky or a real regression?

Check whether the same test passed on master or on rerun within the same PR, and confirm against master history using the ci:insights digest states. A failure sporadic across multiple branches over more than a day is classified as flaky.

Can this Skill rerun a failed CI job or post a comment on GitHub?

No. Rerunning or canceling runs, posting comments, pushing commits, editing workflows, and merging PRs are all forbidden without explicit user approval. The Skill performs read-only inspection and reporting only.

What should I check when many CI jobs fail at once?

Rule out a platform outage first by checking githubstatus.com and status.depot.dev before reading any logs. A burst of runs failing together usually indicates one shared cause such as a bad commit or a GitHub dispatch overflow, not several independent bugs.