Executing Plans

When you've got a written implementation plan and need to execute it methodically, this skill walks you through the process with built-in safety checks. It forces you to review the plan critically before starting, run complexity checks before non-trivial edits, and stop execution if you hit blockers instead of guessing your way through. The structure is rigid by design: load plan, verify each task against budget limits, update checkpoints as you go, then hand off to the finishing skill when done. Works best with subagent support, and it will tell you that upfront. Think of it as guardrails for turning a detailed plan into working code without scope creep or silent overruns.

GitHub stars: 492
last updated: Jun 11, 2026

View on GitHub

Install

$ npx -y skills add ganyuanran/aegis --skill executing-plans --agent claude-code

Installs into .claude/skills of the current project.

Files

Executing Plans

Overview

Load plan, review critically, execute all tasks, report when complete.

Announce at start: "I'm using the executing-plans skill to implement this plan."

For non-trivial execution, include Aegis Visibility: briefly tie the active slice to its plan, checkpoint, drift or verification boundary. At completion, pass plan adherence, evidence, complexity and residual risk to verification-before-completion for the unified receipt.

If subagents are available and the plan has genuinely independent tasks, prefer subagent-driven-development; lack of subagent support does not block inline execution. Same-task agents share one workspace, and the coordinator remains the only Git mutation owner.