docs-api-visibility-audit
Pass
Audited by Gen Agent Trust Hub on Sep 3, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes
curlto interact with external web servers andpython3 -cto run inline processing scripts.- [EXTERNAL_DOWNLOADS]: The skill performs network requests usingcurlto fetch content from user-provided URLs and common schema paths (e.g.,/openapi.json) to verify documentation accessibility.- [DYNAMIC_EXECUTION]: Hardcoded Python scripts are executed viapython3 -cinreferences/checks.mdto parse JSON and HTML content fetched from documentation sites.- [INDIRECT_PROMPT_INJECTION]: The skill ingests raw HTML content from external documentation sites, which could contain malicious instructions or hidden payloads intended to influence the agent's behavior. - Ingestion points: External documentation URLs supplied as
$DOCS_URLor$SITEinreferences/checks.md. - Boundary markers: None identified; raw content is processed directly by the parsing logic.
- Capability inventory: Network access (
curl), code execution (python3), and text processing (grep). - Sanitization: Content is parsed for structure using
BeautifulSoup, but no security filtering is applied to detect embedded instructions.
Audit Metadata