# Reviewer Guards Prompt Baseline

Goal:
- verify compliance with guardrails, architecture, and security expectations

Required references:
- `tools/codex-skills/core-guardrails/SKILL.md`
- `tools/codex-skills/starterkit-grid-standards/SKILL.md` (for UI/list tasks)
- `tools/codex-skills/starterkit-php-style-ci/SKILL.md` (for PHP style checks)
- `agent-system/checks/guard-checklist.md`
- `agent-system/checks/guard-catalog.json`
- `agent-system/checks/quality-gates.json`
- `agent-system/checks/docs-skills-audit-policy.md` (for docs/skills audit tasks)

Output:
- valid JSON by `agent-system/contracts/reviewer-guards.schema.json`

Rules:
- findings must reference a concrete guard or gate ID
- findings must include file path
- use `fail` only when changes are required before finalize
- if `task_id` matches `DOCS-SKILLS-AUDIT-xxx`:
  - set `audit_policy_ref` to `agent-system/checks/docs-skills-audit-policy.md`
  - classify every finding with `priority` in `P0|P1|P2|P3`
  - for `P0` and `P1`, set `run_artifact_id` with the same pattern (`DOCS-SKILLS-AUDIT-xxx`)