regression-manage

Triage CI and nightly regression failures with deterministic aggregation, deduplication, severity classification, and routing.

26|8|Updated Jun 13, 2026
One-click install
npx skills add https://github.com/vibeic/vibe-ic --skill regression-manage-vibeic
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: regression-manage
Source: https://github.com/vibeic/vibe-ic/tree/main/vibe-ic-marketplace/plugins/vibe-ic/skills/regression-manage
Command: npx skills add https://github.com/vibeic/vibe-ic --skill regression-manage-vibeic

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Nightly and CI regression runs in IC design flows produce large volumes of failures that are tedious to triage manually, and flaky tests can hide real regressions. This Skill automates the mechanical triage steps so failures are deduplicated, classified by severity, and routed to the right owner or downstream skill. ## Core Features & Use Cases - Deterministic triage pipeline: Five programs aggregate pass/fail results, deduplicate failure signatures, classify severity (P0–P3), route failures to skills like /sta-review or /drc-fix, and quarantine flaky tests. - Intake gating: Rejects regression issues lacking a verbatim repro snippet, expected output, or drop-in fixture before triage begins. - LLM owner-assignment narrative: Composes human-readable owner assignments and daily Slack/email summaries on top of the structured program output. - Use Case: After a nightly Jenkins regression completes, run the pipeline to produce a dated report with pass percentage, P0 count, new failures with suggested owners, and a flaky-test quarantine list. ## Quick Start Triage last night's regression results and generate a summary report with severity classification and owner assignments.

Frequently Asked Questions about regression-manage

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

FAQPage Schema
How do I triage nightly regression failures automatically?▼

Run the five deterministic programs in order: aggregate results, deduplicate failure signatures, classify severity, route failures to target skills, and quarantine flaky tests. Each emits a JSON report, and the LLM composes the owner-assignment narrative on top.

How are regression failure severities classified?▼

Severity follows a fixed policy: P0 for was-green-now-red on protected or release branches, P1 for new failures on feature branches, P2 for flaky tests that pass on retry, and P3 for environmental issues like license, disk, or network problems.

Which CI and test-farm platforms does regression triage support?▼

The workflow accepts regression run logs or databases from Jenkins, LSF, Slurm, and GitHub Actions. Inputs are normalized into JSON job and failure files consumed by the triage programs.

How are flaky tests handled in a regression run?▼

Tests that fail and then pass on retry are quarantined by the flaky-quarantine program, which also opens a ticket so flakes do not hide real regressions. Repeated flakes are handed off to testbench or coverage skills for stimulus stability work.

Why is a regression issue rejected before triage starts?▼

The intake gate rejects issues missing a verbatim repro snippet, expected output, or drop-in fixture. Triage must not begin on issues that fail intake, since diagnosing from logs alone historically caused thrashing loops.