The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external content from a 'QA Runbook' to generate developer guides.
Ingestion points: The skill reads data from task descriptions, attachments, and the QA Runbook via the sociuu-task tool, and environment status via sociuu-dock.
Boundary markers: There are no explicit instructions or delimiters defined to help the agent distinguish between its own system instructions and potentially malicious instructions embedded within the untrusted QA Runbook data.
Capability inventory: The skill has the capability to execute shell commands (via sociuu-dock) and write data back to task management systems (via sociuu-task).
Sanitization: The instructions do not include any steps for sanitizing, escaping, or validating the input data retrieved from the external runbooks before it is used to shape the guide or stored back in the task system.

sociuu-qa-guide