What problem does it solve? Indexing many repositories one by one is slow and repetitive. This Skill scans the .claude/repo directory, finds repositories missing PROJECT_INDEX.md or PROJECT_INDEX.json, and generates both index files for all of them concurrently using parallel subagents. ## Core Features & Use Cases - Batch Repository Scanning: Lists all repositories under .claude/repo using direct filesystem tools (ls, Glob), excluding the _read directory and bypassing git-index limitations caused by .gitignore. - Parallel Subagent Indexing: Launches one background general-purpose subagent per repository in a single message, each loading the sc:index-repo skill to analyze structure, entry points, and core modules. - Deterministic Verification: Validates results by checking file existence and byte size with wc instead of reading subagent transcripts, then reports total and missing index counts. - Use Case: After cloning ten project repositories into .claude/repo, run this Skill once to produce human-readable PROJECT_INDEX.md and machine-readable PROJECT_INDEX.json for every repository in parallel. ## Quick Start Ask the AI to create PROJECT_INDEX.md and PROJECT_INDEX.json for all repositories under .claude/repo that do not have indexes yet, running each repository in a parallel subagent.