Convert change requests into scoped plans, task lists, and context files.
npx skills add https://github.com/conn-castle/agent-layer --skill plan-work-conn-castle
Please help me install this Agent Skill. Skill: plan-work Source: https://github.com/conn-castle/agent-layer/tree/main/internal/templates/skills/plan-work Command: npx skills add https://github.com/conn-castle/agent-layer --skill plan-work-conn-castle
This skill helps teams convert vague change requests into concrete, auditable plans and task lists, reducing guesswork and rework during project execution.
Provide a plain-language request and optional targets; the skill will return three artifacts: a narrative plan, a task list, and a context file.
High-intent search queries and answers about installing and using this skill.
To convert a vague software change request into an actionable plan, supply a plain-language description and optional targets. The skill outputs a narrative plan, a task list, and a context document scoped specifically for agent execution.
The best way to plan an architectural change for an AI agent is to generate a dedicated context document alongside a narrative plan and task list. This scopes boundaries and validation criteria clearly, preventing scope creep during automated execution.
Yes, you can use this approach for roadmap slices, feature requests, and design artifacts. It converts these change types into a scoped execution path with a narrative plan, task list, and context file.
Prevent scope creep by generating a scoped context document that defines clear boundaries and validation criteria alongside the task list. This ensures the agent executes only the planned architectural changes without expanding the project scope.
No, providing specific file targets is optional when creating a software engineering plan. You can supply a plain-language request alone, and the skill will generate the narrative plan, task list, and context file automatically.
Converting a change request generates three markdown artifacts: a narrative plan file, a compact task list file, and a context document file. These are saved under a temporary run-id directory for immediate agent access.
hulusi-tuncRewrite complex text into plain language with before-and-after comparisons.
EpiphyticExecute implementation plans through structured review, batching, and verification workflows.
ddobrinRefine architectural plans by applying user feedback to Markdown documents.
bllshttngGenerate implementation blueprints with task lists and dependency checks from design documents.
