What makes a plan 'agent-oriented' versus human-oriented?

Agent-oriented plans prioritize actionable 'what' over explanatory 'why,' use structured markdown tags like and , and focus on extreme conciseness to prevent LLM context rot.

How many iterations does the gap-analysis process perform?

The skill can run up to 5 iterations per phase to ensure that all identified gaps are fixed and verified before finalizing the plan.

Does Plan Fixer modify my original documentation files?

No, it follows a non-destructive workflow. It reads your source file and generates a new .PLAN.md file in the same directory, preserving your original human-readable notes.

When should I use the 'split-only' parameter?

Use 'split-only' when you have already manually optimized a plan and just need to generate the .long-run/ directory structure and STATE.md files for an orchestrator.

How does Codex cross-validation improve the plan?

In Phase 2, the skill calls a Codex bridge to review the plan for technical completeness, identifying edge cases or architectural gaps that the primary model might have overlooked.

Plan Fixer

Name: Plan Fixer
Author: fiatkongen

byfiatkongen

Productivity & Workflow

Optimizes implementation plans for autonomous coding agents through iterative gap analysis and structural refinement.

About

Plan Fixer is a specialized Claude Code skill designed to transform human-readable project specifications into lean, instruction-heavy plans optimized for autonomous agent execution. By employing a multi-phase workflow—including automated gap analysis, Codex cross-validation, and automatic plan splitting—it ensures implementation plans are complete, concise, and structured to prevent context rot. This skill is essential for developers using orchestrators who need to ensure their AI agents have clear, actionable, and verified roadmaps before beginning complex coding tasks.

Key Features

Converts human-oriented prose into structured, agent-ready command blocks.
Cross-validates plans with Codex to ensure technical accuracy and completeness.
Maintains non-destructive workflows by generating sidecar PLAN.md files and mapping tables.
Performs iterative gap analysis to identify missing instructions or unhandled edge cases.
Automatically splits complex plans into manageable task groups for long-run execution.
0 GitHub stars

Use Cases

Splitting massive development plans into individual task modules for better orchestrator performance.
Verifying a complex implementation plan before starting a long-running autonomous task.
Converting vague project requirements into actionable, structured agent instructions.

About

Key Features

Converts human-oriented prose into structured, agent-ready command blocks.
Cross-validates plans with Codex to ensure technical accuracy and completeness.
Maintains non-destructive workflows by generating sidecar PLAN.md files and mapping tables.
Performs iterative gap analysis to identify missing instructions or unhandled edge cases.
Automatically splits complex plans into manageable task groups for long-run execution.
0 GitHub stars

Use Cases

Splitting massive development plans into individual task modules for better orchestrator performance.
Verifying a complex implementation plan before starting a long-running autonomous task.
Converting vague project requirements into actionable, structured agent instructions.