Tools provided
Tools the agent exposes — your AI client calls them automatically when it needs them.
- run_full — Orchestrated path: run the full coverage pipeline over the per-screen inventory you provide and return the report — per-screen coverage, one-offs, repeated deviations, merge suggestions, KPIs and a narrative.
- analyze — Run pure deterministic coverage analysis over a provided per-screen inventory with no model call and return scored per-screen findings.
- compute_screen_coverage — Return per-screen design-system coverage — design-system components as a share of total instances — sorted worst-first.
- list_one_offs — Return a flat list of every one-off, non-design-system component instance across all screens.
- top_deviations — Return the one-off components repeated on two or more screens — the deviations that signal a missing design-system component rather than a local exception.
- merge_suggestions — Return the promotion list: recurring deviations worth folding into the design system, and heavily repeated single-screen one-offs, each with the number of screens it would fix.
- overall_coverage — Return the overall design-system usage ratio across the whole provided surface — the single number for a review.
- list_capabilities — List the agent's static capabilities: the available tools and a short description. Useful for discovery.
- plan_inputs — Plan the inputs for a tool: returns the clarifying questions, the JSON schema and a ready-to-edit example.
- discover_intent — Understand your goal and co-design the exact input through clarifying questions before running.
- open_form — Opens the guided form to collect the per-screen component inventory. The entry point for running the agent on your own data.
- run_form — Internal: invoked by the guided form when the user submits. Not for direct use.
