How does this skill handle failing tests during implementation?

The skill includes a verification iteration pattern that analyzes error output, identifies the failing component, attempts to fix the issue, and re-runs tests up to three times before marking a task as blocked.

How does this skill integrate with Plan Marshall?

It is invoked by the task-execute-agent and utilizes the manage-tasks script notations to read task context, update step statuses, and record execution summaries.

What is the two-tier skill loading pattern?

It is an architecture where the agent first loads system-level workflow skills and then dynamically loads domain-specific skills (like Java or React) based on the specific task requirements.

Can this skill work with any programming language?

Yes, it is domain-agnostic. It defines the 'how' of the workflow while applying specific code patterns from domain-specific skills (e.g., pm-dev-java) listed in the task's metadata.

What happens if a task has missing dependencies?

The skill analyzes the file dependency chain; if a dependency is in a later step, it can reorder the workflow or create a minimal stub to ensure implementation continues smoothly.

AI Task Implementation Workflow

Name: AI Task Implementation Workflow
Author: cuioss

bycuioss

生産性とワークフロー

Standardizes the end-to-end execution of coding tasks through a structured understand-plan-implement-verify lifecycle.

概要

This skill provides a domain-agnostic framework for executing implementation tasks within the Plan Marshall orchestration layer for Claude Code. It ensures consistency by enforcing a strict workflow that includes deep context gathering, iterative planning, code modification based on domain-specific patterns, and automated verification. By utilizing a two-tier skill loading system, it combines general implementation logic with specific domain expertise (such as Java or JavaScript) to deliver reliable, production-ready code while recording lessons to prevent repetitive errors and improve future performance.

主な機能

Integrated lesson recording for continuous learning and observation tracking
Two-tier skill loading for combining system workflows with domain-specific expertise
Iterative verification with up to three automated self-correction attempts
Domain-agnostic execution compatible with various tech stacks and project conventions
Automated task state management and real-time progress tracking
0 GitHub stars

ユースケース

Automating the implementation of planned features within complex codebases
Ensuring consistent coding patterns across polyglot microservices
Self-healing task execution that fixes unit test failures automatically

概要

主な機能

Integrated lesson recording for continuous learning and observation tracking
Two-tier skill loading for combining system workflows with domain-specific expertise
Iterative verification with up to three automated self-correction attempts
Domain-agnostic execution compatible with various tech stacks and project conventions
Automated task state management and real-time progress tracking
0 GitHub stars

ユースケース

Automating the implementation of planned features within complex codebases
Ensuring consistent coding patterns across polyglot microservices
Self-healing task execution that fixes unit test failures automatically