Skip to content
Back to skills

Vcsdd Adversarial Refinement

ASecurity

Use this skill when running or interpreting VCSDD adversarial reviews. Provides calibration guidance for the Adversary agent, finding severity classification, and anti-leniency enforcement patterns.

  • 27 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added June 1, 2026
code-qualitygotestingrefactoringsecurity

Security analysis

A100/100

Scanned June 1, 2026

npx -y skills add sc30gsw/vcsdd-claude-code --skill vcsdd-adversarial-refinement --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Vcsdd Adversarial Refinement?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Vcsdd Adversarial Refinement
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/sc30gsw-vcsdd-adversarial-refinement/badge)](https://www.skillsdirectory.com/skills/sc30gsw-vcsdd-adversarial-refinement)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: vcsdd-adversarial-refinement
description: Use this skill when running or interpreting VCSDD adversarial reviews. Provides calibration guidance for the Adversary agent, finding severity classification, and anti-leniency enforcement patterns.
origin: VCSDD
---

# VCSDD Adversarial Refinement

## When to Activate
- Phase 3 (adversarial review)
- Interpreting adversary findings
- Calibrating review intensity

## Adversary Calibration Principles

The Adversary MUST:
1. **Never** say "overall looks good" or equivalent
2. **Always** cite real file paths and line numbers
3. **Always** produce a binary PASS or FAIL per dimension
4. **Never** soften findings with qualifiers like "minor" or "nitpick"
5. **Always** specify `routeToPhase` for actionable routing

## Severity Classification

| Severity | Definition | Examples |
|----------|------------|---------|
| critical | Pipeline cannot continue | Missing spec requirement entirely, security vulnerability |
| high | Significant defect requiring rework | Wrong return type, missing error handling, untested edge case |
| medium | Defect that reduces quality | Misleading name, unnecessary abstraction, dead code |
| low | Improvement opportunity | Style inconsistency, minor duplication |

## Common Leniency Traps

Watch for these patterns and reject them:
- "The code handles most cases" → HOW MANY? WHICH ONES MISSING?
- "Error handling is adequate" → ADEQUATE BY WHAT STANDARD?
- "Generally follows the spec" → WHICH REQUIREMENTS ARE MISSING?
- "Minor refactoring could help" → IS THIS STRUCTURAL_INTEGRITY FAIL OR NOT?

## Finding Routing Guide

| Category | Meaning | Route To |
|----------|---------|----------|
| spec_ambiguity | Requirement or contract wording is ambiguous | 1a |
| spec_gap | Behavior exists but spec/contract does not cover it | 1a |
| requirement_mismatch | Spec is clear, but tests or implementation do not satisfy it | 2b |
| missing_edge_case | Edge case absent from spec or tests | 1a or 2a |
| test_coverage | Required behavior has no meaningful test | 2a |
| test_quality | Tautological or over-mocked test, or implementation-detail assertion | 2a |
| implementation_bug | Logic defect | 2b |
| error_handling | Incorrect or missing error path | 2b |
| security_surface | Validation/auth/injection issue | 2b |
| code_structure | Cohesion/abstraction/organization issue | 2c |
| naming | Misleading or imprecise names | 2c |
| duplication | Repeated logic that should be unified | 2c |
| proof_gap | Missing proof harness or unverifiable claim | 5 |
| invariant_violation | Proof/testing evidence contradicts required invariant | 5 |
| purity_boundary | Supposedly pure logic still depends on effects | 2c or 5 |

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.