Audit → GitHub Issues Publisher

Use the Claude Code command as the canonical workflow instead of maintaining a duplicated snapshot.

Load the Canonical Command

1. Read .claude/commands/audit-publish/SKILL.md completely.
2. Read .claude/commands/_audit-common.md and .claude/commands/_audit-severity.md completely before publishing.
3. Follow the canonical command's report parsing, path-validation gate, status filter, per-finding validation, dedup, label reconciliation, issue creation, snapshot, and completeness-gate steps.

Adapt Claude Conventions to Codex

• Treat the user's text after the skill request as the command arguments described by the canonical skill's argument-hint (the path to the audit report).
• Resolve each /audit-<name> reference to that audit's SKILL.md under .claude/commands/, treating them as pointers that travel into issue bodies — publishing validates findings itself and does not run sibling audits; treat /fix-issue mentions as user-facing next steps only.
• Prefer the repository's codebase-memory graph tools for re-mapping moved Location: paths to their current submodules; use text search for the canonical grep -rn symbol checks against current code.
• Interpret Claude Task-agent instructions as Codex sub-agent delegation only when delegation is available and allowed. Otherwise execute the publish steps in canonical order.
• Creating the GitHub issues is this skill's deliverable: file every CONFIRMED NEW finding via gh issue create, while keeping the canonical status filter, open-issue dedup, label reconciliation against the live repo, and the terminal completeness gate so no finding is double-filed or silently skipped.

Do not copy the canonical command body into this skill. The Claude command is the single source of truth.