ai-ready-fixlisted

# AI Ready Fix Correct proven gaps in agent guidance while preserving project decisions. Aim for score 5 under the `ai-ready-score` rubric, without changing scope or inventing rules to obtain the score. ## 1. Establish the initial assessment Read the `ai-ready-score` skill available in the environment in full and run its workflow on the target repository. It is the authority for the rubric and report; locate it in the catalog if needed. If unavailable, report the dependency and do not promise a validated score. Record the initial score, version, tools in scope, and findings by ID. Even with an existing report, check the current file state before planning fixes. Check local changes and read existing instructions before replacing them. The result of this step is a list of findings that are still valid, each with evidence and affected criterion. If a confirmed score of 5 already exists, finish without rewriting files. ## 2. Resolve knowledge gaps Explore only what the findings require: manifests/runtime, modules, commands and CI, code conventions, tests, and existing rules. Cite the paths supporting each conclusion. In large projects, use subagents for independent topics when available; in small projects or without delegation, perform the same investigation directly. Do not depend on specific tool names such as `Explore` or `AskUserQuestion`. For each finding, define the smallest fix and its verification. Preserve valid specific rules, subtree differences, and tool adapta