FreeWorkflowLow riskv1.1.0

prompt-regression-tester

Run structured prompt regression checks before shipping agent changes.

Defines fixed cases, expected properties, and regression gates for prompt and agent workflow changes.

AI Engineering #prompts #regression #evals #quality

Browse topics and task labels

Prompt or reference Build Edit Redesign Critique Design Review Text

Publisher: skillforge/ai-engineering
Source: Kairaxis manifest
Category: AI Engineering

Package validated
Benchmark tested

Overview

Who it is for: Teams changing prompts, tools, or agent graphs frequently.
Not intended for: One-off creative prompting with no quality bar.
Export formats: codex_skill, markdown_export

Sandbox comparison

Generic baseline vs prompt-regression-tester on the same input, pinned to v1.1.0.

Sample input

gpt-4.1-mini · temp 0.2

System prompt rewrite improved tone but broke refund policy refusals.

Generic baseline

40/100

Looks friendlier. Ship it and monitor.

Skill-enhanced

92/100

Regression report:
- 12/15 cases pass
- Failures: refund refusal, escalations, PII handling
Gate: block ship until refusal cases pass; keep tone cases as non-blocking signals.

Rubric

+52 lift

Coverage of failure modes2 5

Actionable recommendations3 5

Measurable next steps2 4

Risk / limitation awareness2 4

Assumptions

Sample input is representative of a real production task
Same model and temperature for baseline and skill-enhanced runs

Runs on registry sample cases through the shared, deterministic sandbox engine.

Usage

01Review the quality signals, limitations, and permissions on this page.
02Inspect upstream content and license before loading a package.
03Apply relevant guidance within your requested scope and verify the result.
04Agents can also read the machine-readable contract at /api/v1/skills/prompt-regression-tester/manifest.

npx skills add skillforge/ai-engineering --skill prompt-regression-tester

Example prompts

Requests that put this skill to work.

Create a regression suite for our support agent prompt
Diff these two system prompts against critical cases

Limitations

Property checks still need human-authored expected behaviors.

Version history

v1.1.0CurrentJul 1, 2026
Flaky-case quarantine guidance
v1.0.0May 20, 2026
Initial regression pack

At a glance

Installs: 31.2K
8-week activity: 14.1K
Evaluation: Registry example only
Last updated: Jul 1, 2026
Version: 1.1.0

Evidence

Source metadata and registry examples. Content and outcome quality require review.

Package validated
Benchmark tested
Taste calibrated
Creator reviewed

Permissions

Read filesNot needed
Write filesNot needed
Web accessNot needed
API accessNot needed
MCP accessNot needed
Code executionNot needed
External actionsNot needed

Environment fit

Runtime compatibility has not been verified.

Resources

Own this skill?

It is listed under skillforge/ai-engineering. Claim it to manage the listing after review.

False claims may result in permanent bans. Ownership transfer is never automatic.

prompt-regression-tester

Run structured prompt regression checks before shipping agent changes.

Defines fixed cases, expected properties, and regression gates for prompt and agent workflow changes.

Publisher

skillforge/ai-engineering

Source

Kairaxis manifest

Overview

Generic baseline

Skill-enhanced

Rubric

Assumptions

Usage

Example prompts

Limitations

Version history

Related workflows and skills

rag-evaluation-checklist

agent-observability-reviewer

find-skills

Overview

Generic baseline

Skill-enhanced

Rubric

Assumptions

Usage

Example prompts

Limitations

Version history

Related workflows and skills

rag-evaluation-checklist

agent-observability-reviewer

find-skills