From plan-using-file
This skill should be used when the user asks to "update the plan", "update the plan file", "review the plan", "process the plan file", "read plan.md", "check my plan comments", "sync the plan", "plan-using-file", or wants to iterate on a plan using a markdown file as shared state.
How this skill is triggered — by the user, by Claude, or both
Slash command
/plan-using-file:plan-using-fileThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Use a markdown file as mutable shared state for iterative planning. The user edits the file directly — adding questions, comments, and corrections — then invokes this skill to have Claude process all annotations and update the file in place. Works with any markdown file, not just `plan.md` — the default is just a convenience. This skill is designed to be invoked repeatedly within a session.
Use a markdown file as mutable shared state for iterative planning. The user edits the file directly — adding questions, comments, and corrections — then invokes this skill to have Claude process all annotations and update the file in place. Works with any markdown file, not just plan.md — the default is just a convenience. This skill is designed to be invoked repeatedly within a session.
EnterPlanMode or ExitPlanMode.plan.md in the project root.# Plan: [Title]
## Goals
## Approach
## Phases
### Phase 1
### Phase 2
## Open Questions
The user communicates through CriticMarkup comment syntax:
| Marker | Meaning |
|---|---|
{>> comment <<} | User question or comment to address |
{>> @reply: response <<} | Claude's response (persisted for context, removable by user) |
{>> @done <<} | Resolved thread — remove on next pass |
Important: Ignore any {>> ... <<}, {>> @reply: ... <<}, or {>> @done <<} markers that appear inside fenced code blocks or inline code spans — they are examples or documentation, not actionable annotations.
Read the plan file, then process it following these rules:
{>> ... <<} annotation. Do not skip any.{>> @reply: ... <<} immediately after the {>> ... <<} so the user can review and mark {>> @done <<} when satisfied.{>> @done <<} marker, remove the entire resolved thread — the originating {>> ... <<}, any associated {>> @reply: ... <<} annotations, and the {>> @done <<} marker itself. If a {>> @done <<} appears without a preceding {>> ... <<}, remove just the {>> @done <<} and any {>> @reply: ... <<} immediately adjacent to it.If this was the first pass (skeleton just created), ask the user to fill in the plan and return — do not output the standard summary or reminder.
Otherwise, after writing the file and summarizing changes, end your response with a brief reminder such as:
"Review the changes, add any
{>> comment <<}annotations, and run/plan-using-filewhen ready."
If the file has no {>> ... <<} annotations, review the plan and report its current state without making changes. Summarize what the plan covers and note any sections that seem incomplete or worth discussing.
npx claudepluginhub dmose/plan-using-fileUses persistent markdown files (task_plan.md, findings.md, progress.md) as working memory on disk, helping Claude maintain context across complex tasks by writing down decisions, discoveries, and progress.
Creates persistent markdown files (task_plan.md, notes.md, deliverable.md) for multi-step work that requires planning, progress tracking, and durable notes across tool calls.