The Agent Skills Directory

[PROMPT_INJECTION]: The skill requires the agent to ingest and analyze external data, such as private application logs and internal source code. This represents a surface for indirect prompt injection, where instructions embedded in the analyzed data could potentially influence agent behavior.
Ingestion points: Private application logs and internal codebases as defined in the Stage 2 discovery steps.
Boundary markers: The instructions do not specify the use of delimiters or warnings to the agent to disregard instructions found within the audited content.
Capability inventory: The skill has access to shell execution via grep and network communication via the requests library.
Sanitization: No explicit validation or filtering of the content within the logs or code is performed before the agent processes it.
[COMMAND_EXECUTION]: The skill utilizes a grep command with a static regular expression to search the local filesystem for API endpoint patterns. This is a standard auditing operation.
[EXTERNAL_DOWNLOADS]: The skill performs network requests to HubSpot's official API domain (hubapi.com) to retrieve account usage statistics. This interaction with a well-known service is expected for the skill's functional purpose.

audit-api-usage