Orchestrate adversarial planning workflows for complex software engineering tasks.
npx skills add https://github.com/alicicek/tale-mode --skill plan-phase-alicicek
Please help me install this Agent Skill. Skill: plan-phase Source: https://github.com/alicicek/tale-mode/tree/main/plugins/tale-mode/skills/plan-phase Command: npx skills add https://github.com/alicicek/tale-mode --skill plan-phase-alicicek
This skill addresses the tendency of AI models to cut corners on complex tasks by enforcing a rigorous, multi-step planning process that prioritizes verification, adversarial review, and clear, testable outcomes.
Invoke the plan-phase skill to architect a robust implementation plan for your current multi-session feature request.
High-intent search queries and answers about installing and using this skill.
Planning complex software architecture requires decomposing large features into verifiable, independently-shippable phases. This skill orchestrates a high-integrity workflow enforcing adversarial review and evidence-based decision making across multi-session development cycles.
Adversarial review in software planning uses a fresh-context sub-agent to stress-test your architectural plans for design holes and stale assumptions before execution begins. It ensures your engineering decisions are validated against reality rather than training memory.
Decomposing large features into shippable phases involves breaking multi-session projects into defined segments with specific gates and rollback strategies. This skill mandates reading actual code to cite specific file lines, ensuring each phase is independently testable.
Yes, verifying planning assumptions against a live codebase requires integration with host-specific read tools and sub-agent frameworks. This skill mandates reading actual code and documentation to validate architectural decisions during the planning phase.
Plan-phase operates independently with no declared dependencies or components, but effectively executing its verification-first planning workflow requires integration with host-specific read tools and sub-agent frameworks to validate assumptions against your live codebase.
AI planning fails on complex engineering tasks due to the tendency of AI models to cut corners by relying on training memory. This skill enforces a rigorous, multi-step planning process prioritizing verification, adversarial review, and clear testable outcomes.
Lingtai-AIIndex Xiaomi MiMo live docs and configure API access from key prefixes.
ivanyeorsokay, here's the thing. I'm not going to write your code for you, but I'll gladly give you some constructive criticism on your draft.
BetaSAnalyze code repositories read-only to provide evidence-based insights.
oddlyspacedExecute multi-phase technical plans with agent orchestration and review cycles.
