attribution

Audit tracked markdown for prose restating content an external source owns without a pointer or a stamped record, and convert copies into links, quoted citations, or four-part stamped records. Breadcrumb-first: citations in or near a passage are the first confirm targets; budgeted search runs only when no breadcrumb exists. Findings carry evidence-gated tiers (fingerprint-confirmed, source-fetched-similar, llm-suspected, not-found); only fingerprint-confirmed copies are fix-eligible. Also flags verification stamps past their expiry window. Use when: 'find copied content', 'is this copied from the docs', 'check our docs for copied text', 'replace copies with links', 'find stale verification stamps', 'audit provenance', 'where did this paragraph come from', or before publishing prose that restates an upstream page. Read-only by default; explicit 'fix' applies dispositions behind a semantic-diff guard and live pointer checks, and 'sweep' adds per-file closure. Empty target audits tracked markdown.

Set up and maintain this repository's attribution audit configuration: `.claude/attribution.json` across the config cascade's three layers. Manages the categorical exclusions (including the eval-fixture tree, which is a config entry by design and never a rule in a script), the per-candidate and corpus fetch budgets, the separation-rule constants, the stamp expiry window, the accuracy dials for nomination passes and judge sampling, and the fix-eligibility gates. Enables the off-by-default trigger-less-stamp check for a repository whose stamp forms are uniform enough to greppably support it. Use when: 'set up attribution', 'configure attribution', 'exclude a path from the attribution audit', 'change the stamp expiry window', 'the attribution audit flags too much', 'turn on the trigger-less stamp check', after installing the plugin, or to migrate a leftover pre-rename `.claude/provenance.json`. Writes only the consuming repository's own config, never source.