primevue-audit-troubleshooting

PrimeVue Audit And Troubleshooting

Localize the failure first: component API, import/registration, plugin/setup, theme/CSS, accessibility, application state, data/backend, or MCP/plugin tooling. Inspect the narrowest relevant Vue code and local configuration; keep source-confirmed defects separate from optional style suggestions.

For an API audit, call validate_usage on the narrowest complete failing usage. After a reported issue, call get_component for the rejected prop, event, slot, directive, or pass-through contract, patch minimally, and call validate_usage once on corrected final code. Treat metadata-incomplete or unavailable API data as a validation limit, not proof. Do not call version routinely.

For missing MCP, skill discovery, startup, resources, tools, or validation—and for suspected duplicate direct MCP plus plugin configuration—run:

primeui doctor --json --tool <tool> --library primevue

Use no stale or fallback doctor command. Report check identifiers, details, suggestions, and whether the same PrimeVue MCP is configured directly and through the plugin. Retry MCP work only after a passing repair. If MCP remains unavailable, name any current local generated-doc/package-metadata fallback and never call it MCP-validated.

For product, licensing, access, availability, pricing, support, or release claims, use only current official package/license/product sources; otherwise say the claim is unconfirmed. Never expose credentials or license values.

Scenario Contracts

• audit-invalid-api: validate_usage -> get_component -> validate_usage; maximum 3.
• missing-mcp: none; maximum 0.
• duplicate-mcp-plugin: none; maximum 0.

After static repair, run focused type/build/test/browser checks appropriate to the actual failure and report unverified runtime, visual, accessibility, responsive, and business behavior separately.