Stats
Actions
Tags
From cpr
Start a work session. Reads project memory, presents current state, and asks what to work on.
How this skill is triggered — by the user, by Claude, or both
Slash command
/cpr:startThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Resume work on a project that uses the `.claude/memory/` tracking structure.
Resume work on a project that uses the .claude/memory/ tracking structure.
Read .claude/memory/MEMORY.md. It should already be @-imported in .claude/CLAUDE.md,
but read it explicitly to ensure you have the latest content and the full list of
available topic files.
Always read status.md next. Do NOT read other topic files yet.
Tell the user:
Keep this to a few sentences — not a wall of text.
Use AskUserQuestion to ask what the user wants to focus on this session.
Offer the next priorities from status.md as options, plus a free-form option.
Based on what the user picks, read topic files on demand:
spec.md — if the work involves scope or goal questionsplan.md — if the work involves planning or phase decisions.claude/memory/features/<feature>/ — if working on a specific feature (project mode).claude/memory/tasks/<task> file or directory — if working on a specific task
(workspace mode)decisions.md — if the work involves architectural choiceslessons.md and tools.md — if starting something where past lessons are relevantDo NOT read all files speculatively. Only load what is relevant to the chosen work item.
If the work item is implementation, use TaskCreate to decompose it into concrete
coding steps. Present the task breakdown for approval before starting.
If the work item is research, planning, or spec refinement, proceed conversationally.
npx claudepluginhub torarnv/claude-project-manager --plugin cprMaintains a living STATE.md document tracking current project context, active tasks, blockers, and next steps. Auto-loads at session start or on explicit commands like 'update state' or 'where was I'.
Resumes work on existing projects: lists from registry JSON, reads project_state.md for state summary and phase, scans in-progress tasks, suggests next actions. Triggers: 'pick up where I left off'.
Internal skill that checks git state, Plans.md status, AGENTS.md roles, and retrieves resume context from Unified Harness Memory (harness-mem DB) at session start. Reports oversized files and suggests cleanup.