adversarial / default
OfficialCatch failure modes before shipping
Software Engineering#risk assessment#audit trails#bash safety#adversarial review#code governance#prompt injection mitigation
AuthorCollectively-Gary
Version1.0.0
Installs0
System Documentation
What problem does it solve?
This Skill helps prevent unsafe or low-quality outputs from being shipped by independently stress-testing an action’s plan and draft, then surfacing blockers before approval.
Core Features & Use Cases
- Parallel critic review: Runs four generic critics in parallel to identify failure modes, likely audience misinterpretations, second-order consequences, and issues in any recorded bash invocations.
- Risk-calibrated depth: Increases review thoroughness for high-risk actions by adding domain-specific reviewers if configured for the playbook.
- Ship vs. hold gate: Returns
shipwhen no blockers exist, orholdwith a rework brief when blockers are found. - Durable review persistence: Writes findings to a recoverable stage markdown file for traceability, and can post PR issue comments for PR-based reviews.
Quick Start
Run the adversarial review for a pending action to determine whether it should be held for rework or shipped with notes.
Dependency Matrix
Required Modules
None requiredComponents
Standard packageđź’» Claude Code Installation
Recommended: Let Claude install automatically. Simply copy and paste the text below to Claude Code.
Please help me install this Skill: Name: adversarial / default Download link: https://github.com/Collectively-Gary/GaryOS/archive/main.zip#adversarial-default Please download this .zip file, extract it, and install it in the .claude/skills/ directory.
Agent Skills Search Helper
Install a tiny helper to your Agent, search and equip skill from 471,000+ vetted skills library on demand.