performance-optimizer
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONMETADATA_POISONINGINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses
rg(ripgrep) to search for inefficient patterns such asonMountwithfetchin the project's Svelte files. This is a standard local search operation. - [METADATA_POISONING]: The reference file
references/patterns.mdcontains Next.js/React patterns which are inconsistent with the skill's stated SvelteKit focus, representing a non-malicious metadata error. - [INDIRECT_PROMPT_INJECTION]: The skill identifies a vulnerability surface where codebase content is scanned and processed by the agent. 1. Ingestion points: Raw codebase content scanned by
rg. 2. Boundary markers: None defined in the skill instructions. 3. Capability inventory: File system search and code modification tools. 4. Sanitization: No sanitization logic for ingested code content is specified.
Audit Metadata