exa-hello-world
Pass
Audited by Gen Agent Trust Hub on Sep 12, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill demonstrates best practices for credential management by instructing the user to use the
EXA_API_KEYenvironment variable instead of hardcoding API keys. - [SAFE]: The skill uses the official
exa-jsSDK, which is the standard library for the Exa search service, a well-known neural search provider. - [INDIRECT_PROMPT_INJECTION]: The skill defines a vulnerability surface where untrusted data from the web is ingested via the Exa search API. The risk is minimal because the demonstration code only logs output to the console and does not perform sensitive automated actions with the retrieved data.
- Ingestion points:
SKILL.mdcontains code blocks that fetch search results, highlights, and summaries from external websites using the Exa search API. - Boundary markers: No specific delimiters or instructions to ignore embedded content are present in the provided snippets.
- Capability inventory: The logic is limited to search retrieval and
console.logstatements; no file-writing, subprocess execution, or subsequent network operations are present within the skill's scripts. - Sanitization: No data sanitization or escaping is implemented in the simple demonstration snippets.
Audit Metadata