The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests data from external API responses which is then used in subsequent tool calls, creating a theoretical injection surface.
Ingestion points: API responses from VM_ANOMALY_URL endpoints (SKILL.md).
Boundary markers: No specific delimiters are used in the interpolations.
Capability inventory: The skill uses curl for network requests and jq for data transformation (SKILL.md).
Sanitization: The skill uses jq to construct JSON request bodies, which serves as a form of structured data sanitization.
[DATA_EXPOSURE]: The skill interacts with sensitive configuration data but implements robust protections. It provides jq logic to explicitly redact user credentials from datasource URLs and advises the use of secure configuration files (VM_CURL_CONFIG) for authentication tokens to avoid exposure in the process list.

vmanomaly-query