The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a local Python script (scripts/gdpr_deadlines.py) using the Bash tool to calculate legal deadlines (72h for breaches and 1 month for data subject requests). The script is self-contained, deterministic, and does not have any external dependencies or network operations.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to process incident descriptions provided by users, which serves as an ingestion point for untrusted data.
Ingestion points: User-pasted incident descriptions are processed by the model (SKILL.md).
Boundary markers: None explicitly defined in the instruction text to separate user data from system instructions.
Capability inventory: Access to the Bash tool (limited to the provided deadline calculator script) and the Read tool.
Sanitization: No explicit sanitization or filtering of the incident description is defined.
Risk Assessment: The risk is low as the skill explicitly requires human approval for all actions, does not automate network egress, and limits tool usage to clerical tasks.
[SAFE]: The skill adheres to security best practices by setting pii-egress: none, data-residency: local, and requires-human-approval: true. No hardcoded credentials, persistence mechanisms, or obfuscation techniques were found.

rodo-naruszenie-72h-pl