ck

What problem does it solve? AI coding sessions lose all context when they end, forcing you to re-explain your project's goal, stack, decisions, and progress every time. This Skill maintains a persistent per-project memory so each new session starts with a compact briefing of where you left off. ## Core Features & Use Cases - Project Registration & Auto-Detection: The /ck:init command scans package.json, go.mod, Cargo.toml, pyproject.toml, CLAUDE.md, and README.md to auto-detect your project's name, stack, goal, and repo. - Session Save & Resume: /ck:save records a structured session summary (left-off point, next steps, decisions, blockers, git activity) into context.json, and /ck:resume renders a full bordered briefing box to pick up exactly where you stopped. - SessionStart Hook: A hook injects a ~100-token compact summary on every session start, detecting unsaved sessions, new git commits, and goal mismatches against CLAUDE.md. - Use Case: You work across three repositories during the week. Run /ck:list to see a portfolio table of all projects with staleness indicators, then /ck:resume 2 to get a full briefing on the second project including its last decisions and next steps. ## Quick Start Ask the assistant to run /ck:init to register the current project, then use /ck:save at the end of each work session and /ck:resume when you return.