The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill guides the agent to collect user-provided parameters, such as organization and project identifiers, to perform resource management tasks via tools like harness_create. This creates a surface where user-supplied data is interpolated into tool calls.
Ingestion points: Step 1 and Step 4 in SKILL.md collect organization and project data.
Boundary markers: No explicit delimiters or validation instructions are provided in the instruction set.
Capability inventory: The skill uses harness_list and harness_create tools to interact with the Harness platform.
Sanitization: No specific input sanitization steps are documented in the instructions for these parameters.
[SAFE]: The skill is authored by Harness and facilitates legitimate administrative actions within the Harness ecosystem. No indicators of exfiltration, persistence, or malicious code execution were identified.

manage-artifacts