dd-unblock-pr

Diagnoses failing PR CI pipelines by classifying failures as flaky, infra, or regression.

5|2|Updated Mar 9, 2026
One-click install
npx skills add https://github.com/lauhon/pi --skill dd-unblock-pr-lauhon
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: dd-unblock-pr
Source: https://github.com/lauhon/pi/tree/main/skills/dd-unblock-pr
Command: npx skills add https://github.com/lauhon/pi --skill dd-unblock-pr-lauhon

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? When a pull request's CI pipeline fails, engineers waste time manually digging through job logs to figure out whether the failure is caused by their own changes, a flaky test, or broken infrastructure. This Skill automates that triage by querying Datadog CI Visibility data and attributing each failure to a root cause category with a recommended action. ## Core Features & Use Cases - Failure Classification: Attributes each failing CI job as flaky, infra, regression, or unknown using blame-guard checks against the default branch and blast-radius analysis across branches. - Targeted Remediation: Proposes concrete next steps per classification, including GitHub Actions reruns for transient infra failures and handoff to flaky-test deep investigation. - PR Health Reporting: Fetches code coverage for the branch and produces a structured triage brief summarizing all failures with evidence and confidence levels. - Use Case: A developer pushes a branch and sees three red CI jobs. Instead of reading raw logs, they run this Skill to learn one failure is a known flaky test, one is a pre-existing infra issue on main, and one is a genuine regression introduced by their changes. ## Quick Start Investigate why the CI pipeline is failing on my current PR branch and tell me which failures are flaky, infra, or regressions.

Frequently Asked Questions about dd-unblock-pr

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

FAQPage Schema
How do I find out why my PR CI pipeline is failing?

Run this Skill with your PR branch name and it queries Datadog CI Visibility for failing jobs and tests. It classifies each failure as flaky, infra, or regression using blame-guard checks against the default branch and cross-branch failure counts, then outputs a triage brief with recommended actions.

How to tell if a CI test failure is flaky or a real regression?

The Skill checks whether the test appears in Datadog's active flaky tests registry and whether the same job fails on the default branch or many other branches. A failure unique to your branch is classified as a regression, while cross-branch intermittent failures are classified as flaky.

Does this Skill require Datadog and the pup CLI?

Yes, it requires the pup CLI installed and authenticated against Datadog, provided by the dd-pup skill. All CI event, test, flaky test, and coverage queries run through pup commands, so Datadog CI Visibility instrumentation must be set up for your repository.

Can it automatically retry failed CI jobs?

For transient infrastructure failures on GitHub Actions, it extracts the run ID from the pipeline URL and runs gh run rerun with the --failed flag. For deterministic failures or other CI providers, it directs you to the provider UI instead of retrying.

Why does the coverage or quality section show no data available?

Coverage shows no data when the pup code-coverage branch-summary command returns nothing or errors for that branch. Code quality and security violation counts are not exposed by pup at all, so those lines always report no data available.