ai-ready-scorelisted

# AI Ready Score Assess the quality of the instructions an agent actually receives. The score measures guidance available in the repository, not software quality or guaranteed agent performance. ## 1. Define scope and inventory Resolve the repository path and identify tools in scope from an explicit request or project configuration/documentation. Without sufficient evidence, use the interoperability profile with Claude Code, Copilot, Codex, and OpenCode and state that assumption. Keep all four tools in the report; tools outside scope are not applicable and do not lower the score. A shared file does not count as multiple independent contents. Run `python <skill-dir>/scripts/discover.py <repo>` (read-only), replacing `<skill-dir>` with this skill's absolute directory. The inventory is evidence of files, not proof of loading. Read resident instructions and valid local targets once per source; open skills only as needed to check candidates and references. Inspect settings, imports, and subtree instructions that change effective content. External links are dependencies to record, not content to read automatically. Check manifests, scripts, CI, and samples of relevant modules to validate commands and rules. Distinguish a command found in configuration from a command executed successfully. Do not run build/test or install dependencies just to assign a score. Record excluded directories and read errors; expand the directed search if an exclusion hides project code. If evidence i