The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from alert notes and indicators that are later formatted into Markdown and JSON reports. Without proper sanitization, this input could be used to inject malicious instructions into downstream systems or agent contexts.
Ingestion points: In scripts/soc_utils.py, the AlertTriage.add_note, AlertTriage.add_ioc, and ShiftHandover.add_open_alert methods accept arbitrary string inputs.
Boundary markers: None identified; data is directly interpolated into report templates.
Capability inventory: Analysis of scripts/soc_utils.py confirms no dangerous capabilities such as subprocess execution, file system writes, or network access are present.
Sanitization: Absent; the skill does not perform any escaping, validation, or filtering of user-provided content.

soc-operations