setup-agent-routinglisted

# Setup Agent Routing Configure one thin routing index and consumer tracker/domain documentation for the prepared issue workflow. Reuse the canonical readiness and delivery contracts; setup records where to find them, not another execution procedure. ## Authorized Scope Preserve the user's target and existing write authorization. Read-only/report-only requests stay read-only. Existing approval for this setup covers the requested local files; ask only for consequential missing choices or expanded authority. Model, effort, account, checkout and provider policy remain harness-owned. ## Contract Inputs: - Target repository, existing agent instructions and `docs/agents/` layout. - Live tracker/project fields, labels, and user-approved overrides. Outputs: - One `## Agent skills` index and populated tracker, triage and domain documents. - Present/missing state, canonical resource locations and unresolved setup blockers. Creates/Modifies: - The existing `CLAUDE.md` or `AGENTS.md` routing block, updated in place. - `docs/agents/{issue-tracker,triage-labels,domain}.md`, preserving unrelated content. External Side Effects: - Read-only GitHub inspection and authorized local documentation writes. - No issue labels, board changes, workflows, model configuration or dispatch writes. Confirmation Required: - Ask only for a missing target/authority or consequential choice not settled by session context or repository convention. Prepare drafts before any required approval; do