The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection because it ingests data from external sources and has capabilities to perform destructive or creative actions.
Ingestion points: Data enters the agent context through the outputs of harness_list and harness_get tools when fetching user, group, or service account details (SKILL.md).
Boundary markers: The instructions do not define explicit delimiters or instructions to ignore embedded commands within the processed data.
Capability inventory: The skill provides access to harness_create and harness_delete tools, allowing it to modify the platform configuration based on instructions (SKILL.md).
Sanitization: There is no evidence of sanitization or validation of fields like description or name which could be controlled by an attacker within the Harness platform.

manage-users