ai-adviselisted

# Advise Discoverable wrapper around the `ai-advise` governance advisor: dispatches the agent via the Agent tool, captures findings with severity (`info | warn | concern`), and renders an advisory output. Never blocks. Never modifies code. ## Quick start ``` /ai-advise # default: advise mode on staged changes /ai-advise advise src/auth/ # post-edit advisory, scoped to paths /ai-advise gate # pre-dispatch governance check /ai-advise drift # compare implementation to active decisions ``` ## Workflow Principles applied: §10.6 SDD (the advisory output traces every warning to an active decision or stack standard — no advice without a documented anchor); §10.4 DRY (the agent contract owns the analysis loop — the skill never paraphrases standards inline). 1. **Step 0** — load stack contexts: read `.ai-engineering/manifest.yml` `providers.stacks` and apply `.ai-engineering/overrides/<stack>/conventions.md` so stack-specific standards are in scope when guard analyses changed files. 2. **Detect mode** — first positional argument is `advise` (default), `gate`, or `drift`. Anything else is treated as a path filter and the mode defaults to `advise`. 3. **Dependency preflight** — verify `.agents/agents/ai-advise.md` (or, post-rename, `.agents/agents/ai-advise.md`) is on disk. STOP and report the exact missing path if absent — never paraphrase agent instructions inline. 4. **Dispatch** — in