deco-api-call-dedup
Pass
Audited by Gen Agent Trust Hub on Sep 4, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill documents a
curlcommand for inspecting VTEX Intelligent Search API responses. The target domainvtexcommercestable.com.bris associated with VTEX, a well-known e-commerce service provider. The command uses placeholders for the user's specific account and product queries. - [COMMAND_EXECUTION]: The skill recommends using
grepandripgrep(rg) to audit the local codebase for N+1 performance patterns in loaders and components. These are standard developer tools for static analysis. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to analyze the project's source code, which constitutes an ingestion surface for potentially untrusted data if the codebase contains third-party contributions.
- Ingestion points: Source code files in
src/components/andsrc/sections/identified viagrepcommands. - Boundary markers: The instructions do not specify any delimiters or safety warnings for the agent when processing this code.
- Capability inventory: The skill uses file search and read capabilities to identify and suggest fixes for specific code patterns.
- Sanitization: There is no mention of sanitizing or escaping the analyzed code before it is processed by the agent.
Audit Metadata