MANDATORY: Follow the complete shared lifecycle contract in ${CLAUDE_PLUGIN_ROOT}/skills/shared/run-lifecycle.md with declared skill fluff-analysis.

fluff-analysis

MANDATORY: READ ENTIRE FILE before composing user-facing prose: ${CLAUDE_PLUGIN_ROOT}/skills/shared/readability-style.md.

Characterize review fluff from the synchronized run-log cache in the current repository. This includes suggestions that are not accepted (rejected or deferred to Out-of-Scope) or accepted-but-low-value. The analyzer normalizes every review finding (outcome, reviewer/voter severity, semantic tags), then prints a markdown report and recommendations for tightening the reviewer self-filter and judge (voter) instructions.

This is the standing tool behind the kind of analysis filed as a [Analysis Report] issue: re-run it as the corpus grows to track whether necessity-gate changes (see skills/shared/review-acceptance-rubric.md) move acceptance and findings-per-run.

Usage

/fluff-analysis [--include-in-progress] [--cutoff ISO8601] [--since-version X.Y.Z] [--min-group N] [--log-root DIR] [--out FILE]

Run the Analysis

Call the analyzer via the Bash tool, forwarding any flags, then relay its markdown stdout to the user: