What research is this skill based on?

It is based on the 'ReAct: Synergizing Reasoning and Acting in Language Models' paper (Yao et al., 2022) and modern LangGraph execution patterns for robust AI agent behavior.

What is the Plan-and-Execute methodology?

It is an execution framework where the AI treats a plan as a hypothesis, executing steps one by one and observing results to decide whether to continue, retry, or modify the plan based on the actual outcome.

How does the Building skill handle execution failures?

The skill utilizes an Execute-Observe-Decide loop. If a step fails, it can perform a local repair, substitute the approach, add prerequisite steps, or switch to an entirely different path to reach the goal.

Does it support task recovery?

Yes. The skill maintains a detailed loop state and creates checkpoints after successful steps, allowing the AI to resume or rollback to a stable state if a critical failure occurs.

Adaptive Task Builder

Name: Adaptive Task Builder
Author: twofoldtech-dakota

bytwofoldtech-dakota

Productivity & Workflow

Implements a high-precision Plan-and-Execute methodology for adaptive task execution with intelligent replanning.

About

The Building Skill provides an industrial-grade orchestration layer for AI-assisted development, moving beyond static instruction following to a dynamic 'Execute-Observe-Decide' loop. Inspired by ReAct research and LangGraph execution patterns, this skill enables Claude to treat plans as hypotheses that are tested against reality in real-time. By implementing sophisticated error recovery, automated success verification, and five distinct replanning strategies, it ensures that complex engineering tasks reach completion even when encountering unexpected environmental hurdles or brittle dependencies.

Key Features

0 GitHub stars
Dynamic Plan-and-Execute methodology for adaptive reasoning
Built-in success criteria verification and exponential backoff retries
Five intelligent replanning strategies: local repair, substitution, adding steps, skipping, and alternative paths
Comprehensive execution state management with automatic checkpointing
Automated 'Execute-Observe-Decide' loop for industrial precision

Use Cases

Managing multi-step code migrations where environment variables or dependencies may change
Building robust AI-driven development workflows that require recovery from transient failures
Executing complex refactoring tasks that require continuous validation of side effects

About

Key Features

0 GitHub stars
Dynamic Plan-and-Execute methodology for adaptive reasoning
Built-in success criteria verification and exponential backoff retries
Five intelligent replanning strategies: local repair, substitution, adding steps, skipping, and alternative paths
Comprehensive execution state management with automatic checkpointing
Automated 'Execute-Observe-Decide' loop for industrial precision

Use Cases

Managing multi-step code migrations where environment variables or dependencies may change
Building robust AI-driven development workflows that require recovery from transient failures
Executing complex refactoring tasks that require continuous validation of side effects