Can this skill be used in Team Mode?

Yes, this skill is mode-agnostic and functions identically in both Task Mode and Team Mode, though it uses persistent teammates or fresh tasks differently depending on the phase.

When should I avoid using the Orchestrated Execution skill?

It is not recommended for simple single-file bug fixes, minor copy changes, or tasks that do not have a written specification or 'Definition of Done'.

How does adversarial review differ from standard review?

Adversarial review always spawns a fresh AI task instance with no prior context of the implementation phase, ensuring an unbiased, critical assessment against the original spec.

What is the 4-phase execution loop?

It is a specialized workflow consisting of Implementation, Validation, Adversarial Review, and Commitment phases designed to ensure code meets all specifications before it is finalized.

What happens if a quality gate fails?

If validation or review gates fail, the workflow returns to the implementation phase for fixes, followed by re-validation and a fresh adversarial review.

Orchestrated Execution Loop

Name: Orchestrated Execution Loop
Author: jpeggdev

byjpeggdev

0•

安全与测试

Implements a high-integrity, 4-phase execution cycle to ensure code quality and specification compliance through adversarial review.

This skill provides a structured framework for managing complex coding tasks by decomposing them into atomic work units and processing each through a rigorous four-phase pipeline: Implementation, Validation, Adversarial Review, and Commitment. It prioritizes independent verification over agent self-reporting, utilizing fresh task instances for reviews to eliminate bias and ensuring all changes align strictly with defined specifications, coverage thresholds, and architectural quality gates.

主要功能

01Adversarial Review Gate using fresh task instances to prevent context bias

020 GitHub stars

03Comprehensive pre-flight architecture and security checklists

04Independent orchestrator-led validation of tests, linting, and coverage

05Work unit decomposition with explicit dependency graph management

06Rigorous 4-phase loop: Implement, Validate, Adversarial Review, and Commit

使用场景

01Managing multi-agent orchestration where sub-agent output needs adversarial auditing

02Decomposing complex feature epics into verifiable, atomic work units

03Implementing high-stakes architectural changes requiring multi-gate validation

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add jpeggdev/buildwithjpegg orchestrated-execution

For use in Claude.ai and ChatGPT

主要功能

01Adversarial Review Gate using fresh task instances to prevent context bias

020 GitHub stars

03Comprehensive pre-flight architecture and security checklists

04Independent orchestrator-led validation of tests, linting, and coverage

05Work unit decomposition with explicit dependency graph management

06Rigorous 4-phase loop: Implement, Validate, Adversarial Review, and Commit

使用场景

01Managing multi-agent orchestration where sub-agent output needs adversarial auditing

02Decomposing complex feature epics into verifiable, atomic work units

03Implementing high-stakes architectural changes requiring multi-gate validation

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add jpeggdev/buildwithjpegg orchestrated-execution

For use in Claude.ai and ChatGPT