audit-siri-ai

Scan your App Intents code and App Skills for problems that keep Siri AI from finding, choosing, or running your app's actions in the 27.3 Siri AI Developer Preview.

What This Command Does

Launches the siri-ai-auditor agent. Because it reads your source code, it catches problems that appear only when Siri runs an intent, before you build the app and test it on a device.

What It Checks

1. Search tools – Tools that search for your app's content, which Siri already does through Spotlight
2. Tool metadata – Missing intent and parameter descriptions that Siri's model relies on
3. Entities – Entities with no exposed properties, or that Siri can't find
4. Cold launches – @Dependency values that don't exist when Siri starts your app without its UI
5. Results and errors – Values the system can't see, errors shown as generic failures, and work that can repeat when an intent re-runs
6. Saved shortcuts and availability – Entity identifiers and enum values that change between launches, and tools gated wrongly for earlier OS releases or for watchOS and tvOS targets
7. App Skills – SKILL.md problems that keep every skill from shipping, and bundles left out of the app target

It finishes with a readiness score (READY, NEEDS WORK, or BLOCKED).

Usage

/axiom:audit siri-ai

Related

• siri-ai-auditor – The agent that powers this command
• siri-ai – How to bring your app's actions to Siri AI, and the fix behind each finding
• /axiom:siri – Runs this audit, then builds the app and analyzes its metadata with sirimetadataanalyzer