From codescope
Analyze your codebase to build the knowledge graph, detect conventions, and identify danger zones. Produces AI readiness score and cross-service dependency map for monorepos.
How this skill is triggered — by the user, by Claude, or both
Slash command
/codescope:bootstrapThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Analyzes your codebase to build the knowledge graph, detect conventions, identify danger zones, and compute an AI readiness score.
Analyzes your codebase to build the knowledge graph, detect conventions, identify danger zones, and compute an AI readiness score.
Check prerequisites Call the codescope_status MCP tool. Verify config_exists is true. If false, tell the user to run /codescope:onboard first.
Determine bootstrap mode Check if the user passed --force flag. If not, the orchestrator will auto-detect incremental vs full mode.
Handle --force confirmation (per D-30) If --force is specified, before running the bootstrap pipeline, display a confirmation prompt showing what will be rebuilt and what will be preserved:
Will be rebuilt (deleted and regenerated):
Will be preserved:
Ask the user to confirm ("Proceed with full re-bootstrap?" y/n). If the user cancels, abort without running the pipeline. If confirmed, proceed to step 4.
Run the bootstrap pipeline IMPORTANT: The user's current working directory is the project to analyze, NOT the plugin directory. Before running, capture the user's project root, then use absolute paths:
PROJECT_ROOT=$(pwd)
PLUGIN_DIR=$(dirname "$(dirname "$(readlink -f "$0" 2>/dev/null || echo "$HOME/.claude/plugins/marketplaces/codescope")")")
Then run with explicit --project-root:
cd <plugin-directory> && node --import tsx/esm src/bootstrap/run-bootstrap.ts --project-root <user-project-root> [--force]
Always pass --project-root with the user's actual project directory. Never rely on process.cwd() — it will be the plugin directory after cd'ing to find tsx. This runs the full pipeline:
Read and display results
Read the bootstrap output from .claude/codescope/readiness.md and display the completion summary following this format:
AI Readiness: {grade} ({percent}%)
| Dimension | Score | Delta |
|---|---|---|
| Convention Coverage | {pct}% | {+/-N%} |
| Type Safety | {pct}% | {+/-N%} |
| Test Coverage Proxy | {pct}% | {+/-N%} |
| Import Graph Health | {pct}% | {+/-N%} |
Next: /codescope:orient [task]
--force -- Full re-bootstrap. Shows confirmation before wiping (per D-30): lists what will be rebuilt and what's preserved. User confirms or cancels. Rebuilds all analysis artifacts. Preserves config.yml, conventions-enforced.md, learnings.md.npx claudepluginhub jwadhwa2259/codescope --plugin codescopeRuns static tools like tsc, semgrep, knip to analyze codebase for dead code, test quality, duplicates, complexity, security, architecture. Full TS/JS support; limited Python/Go/Rust. Stores structured reports.
Assesses codebase for AI agent readiness by detecting stacks, monorepos, git setup, and evaluating style, testing, code quality, secrets, and file sizes.