How do I install the Inspect Bdd Steps skill?

Run npx skillmds@latest add prebid/inspect-bdd-steps in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Inspect Bdd Steps safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Inspect Bdd Steps?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Inspect Bdd Steps free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Inspect Bdd Steps?

prebid (@prebid) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Inspect Bdd Steps by prebid — Agent Skill

Name: Inspect Bdd Steps
Author: prebid

Inspect Bdd Steps

Two-pass BDD step assertion completeness inspector. Pass 1 (Sonnet): triage all Then steps — FLAG or PASS. Pass 2 (Opus): deep trace flagged steps with full production context, producing architectural judgment on what the correct assertion should be. Use after writing or modifying BDD step definitions to catch assertion mismatches (steps that claim to verify X but actually only check existence).

prebid 86ae6f6 2.0 KB Updated Sep 17, 2026

File contents

BDD Step Assertion Completeness Inspector

Inspects every BDD step function for semantic completeness: does the function body actually implement what the step text claims?

Usage

/inspect-bdd-steps
/inspect-bdd-steps --pass1-only

What It Does

AST scan: Extracts all @given/@when/@then decorated functions from tests/bdd/steps/
Pass 1 (Sonnet triage): For each Then step, asks: "Is there a HIGH chance this function does NOT implement what the step text claims?"
- PASS: Function plausibly implements its claim
- FLAG: Function likely doesn't (pass body, truthiness-only check, etc.)
Pass 2 (Opus deep trace): For each FLAG, collects production context (schemas, error classes, harness code) and asks Opus to make an architectural judgment about what the correct assertion should be.
Report: Writes .claude/reports/bdd-step-audit-<date>.md with findings grouped by severity (MISSING > WEAK > COSMETIC).

Protocol

Run the inspection script:

python3 .claude/scripts/inspect_bdd_steps.py

Options:

--pass1-only — Skip Pass 2 deep trace (fast triage only)
--steps-dir PATH — Override step definitions directory
--output PATH — Override report output path
--then-only — Only inspect Then steps (default: true)

Review the generated report and use findings to inform beads-58u (assertion mismatch fixes) or create new tasks.

When to Use

After writing new BDD step definitions
After modifying existing step assertions
As a periodic audit (monthly or per-epic)
Before closing BDD-related PRs

---name: inspect-bdd-stepsdescription: Two-pass BDD step assertion completeness inspector. Pass 1 (Sonnet): triage all Then steps — FLAG or PASS. Pass 2 (Opus): deep trace flagged steps with full production context, producing architectural judgment on what the correct assertion should be. Use after writing or modifying BDD step definitions to catch assertion mismatches (steps that claim to verify X but actually only check existence).---# BDD Step Assertion Completeness InspectorInspects every BDD step function for semantic completeness: does the functionbody actually implement what the step text claims?## Usage```/inspect-bdd-steps/inspect-bdd-steps --pass1-only```## What It Does1. **AST scan**: Extracts all `@given`/`@when`/`@then` decorated functions from `tests/bdd/steps/`2. **Pass 1 (Sonnet triage)**: For each Then step, asks: "Is there a HIGH chance this function does NOT implement what the step text claims?" - PASS: Function plausibly implements its claim - FLAG: Function likely doesn't (pass body, truthiness-only check, etc.)3. **Pass 2 (Opus deep trace)**: For each FLAG, collects production context (schemas, error classes, harness code) and asks Opus to make an architectural judgment about what the correct assertion should be.4. **Report**: Writes `.claude/reports/bdd-step-audit-<date>.md` with findings grouped by severity (MISSING > WEAK > COSMETIC).## ProtocolRun the inspection script:```bashpython3 .claude/scripts/inspect_bdd_steps.py```Options:- `--pass1-only` — Skip Pass 2 deep trace (fast triage only)- `--steps-dir PATH` — Override step definitions directory- `--output PATH` — Override report output path- `--then-only` — Only inspect Then steps (default: true)Review the generated report and use findings to inform beads-58u(assertion mismatch fixes) or create new tasks.## When to Use- After writing new BDD step definitions- After modifying existing step assertions- As a periodic audit (monthly or per-epic)- Before closing BDD-related PRs

Frequently asked questions