skills.rest
Gentleman-Programming
Gentleman-ProgrammingOFFICIAL
Explore all skills by this author →
GitHub

gentle-ai-bench

Authors and verifies gentle-ai bench journeys with driven harness execution.

6.1k|707|Updated Feb 27, 2026
Search Tags:#test harness#ci verification#Go testing#gentle-ai#End-to-end testing#bench testing#journey corpus
One-click install
npx skills add https://github.com/Gentleman-Programming/gentle-ai --skill gentle-ai-bench
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: gentle-ai-bench
Source: https://github.com/Gentleman-Programming/gentle-ai/tree/main/internal/assets/skills/gentle-ai-bench
Command: npx skills add https://github.com/Gentleman-Programming/gentle-ai --skill gentle-ai-bench

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It prevents false confidence in bench testing by enforcing that journey changes in the gentle-ai bench corpus are proven through driven harness execution rather than relying on go test ./bench, which only validates corpus declarations.

Core Features & Use Cases

  • Journey Authoring Rules: Enforces unique journey IDs across all journeys_*.go files, mandatory Review: declarations (reviewOptedIn or reviewUntouched), and runnable commands on every execute transition.
  • CI-Faithful Verification: Requires reading the Unit Tests step in .github/workflows/ci.yml and reproducing its exact build and gentle-ai-bench run --binary ... commands, with --only <journey-id> for single journeys.
  • Semantic Change Auditing: When a ratified product semantic changes, greps the corpus for journeys pinning the old behavior so stale assertions do not keep defects green.
  • Use Case: When adding a new bench journey or diagnosing a bench failure in CI, load this Skill to author the journey correctly, run both declaration tests and the driven harness, and include the driven-mode summary line in the PR body.

Quick Start

Load this skill before editing anything under bench/ in gentle-ai, then follow its rules to author the journey and prove it with the driven harness against a locally built binary.

Frequently Asked Questions about gentle-ai-bench

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

FAQPage Schema
How do I verify a gentle-ai bench journey actually runs?▼

Build the bench harness and the product binary, then run the harness against the binary using the exact commands from the Unit Tests step in .github/workflows/ci.yml. Use --only <journey-id> to drive a single journey; go test ./bench alone proves nothing about execution.

Why does go test ./bench pass but the journey still fails?▼

go test ./bench only validates corpus declarations such as ID uniqueness and mandatory fields; it never executes journeys. A green run says nothing about driven behavior, so you must run the harness against a locally built binary for real proof.

What is the Review declaration in a bench journey?▼

Every journey must declare Review as either reviewOptedIn, where the runner enables receipt-driven development globally before the first step, or reviewUntouched, when the journey concerns the switch itself or is unrelated to reviews. validateCorpus fails the run without it.

How do I add a new journey without ID collisions?▼

Journey IDs must be unique across every journeys_*.go file, and the collision guard fails loudly naming both files. Read the corpus to pick an unused ID, never reuse a retired one, and prefer a new journeys_*.go file when shared ones are owned by open PRs.

What should I do when a product semantic changes?▼

Grep the journey corpus for journeys pinning the old behavior before shipping, since the corpus is a second test surface beyond unit tests. A journey asserting the old defect keeps that defect green, so update stale pins in the same change.

Related Skills

Explore More Software Engineering Skills→
llmh333llmh333

golang-testing

Create Go unit tests with table-driven cases and subtests.

Community
Intermediate
KaiserWhoLearnsKaiserWhoLearns

analyze-ci

Analyze failed GitHub Action jobs and summarize root causes from PR URLs.

Community
Intermediate
dotBeepsdotBeeps

go-testing

Guide Go tests, benchmarks, and fuzzing with testify and go test flags.

Community
Intermediate
emergent-companyemergent-company
1

run-e2e-test

Execute Go-based end-to-end tests for the emergent.memory project.

Official
Advanced