What problem does it solve? When starting work on a code project with CALM, the AI has no persistent context about the project's identity, decisions, or ongoing tasks. This Skill bootstraps a structured knowledge frame (domain tag, topic tags, decisions, activities, reference pointers) so that future sessions can recover context via check_in and search. ## Core Features & Use Cases - Existing-frame detection: Searches domain tags and activities across all statuses to avoid duplicating or conflicting with an already-established project frame. - Markdown scanning plus interview: Extracts frame candidates from repository docs (README, ADR, CONTRIBUTING) and fills remaining gaps through conversational questioning, with user arbitration on each decision. - Structured two-phase writes: Creates a setup topic carrying the domain and plain tags, then records logs, decisions, and activities, and finally attaches tag notes and descriptions via update_tag. - Use Case: An engineer opens a new repository and asks to set up the project; the Skill scans the README, proposes a domain tag and two topic tags, records the chosen tech stack as a decision, and registers the next task as an activity. ## Quick Start Ask the AI to run coding-project-setup for this repository so it scans the docs and builds the project's knowledge frame.