The Agent Skills Directory

[Indirect Prompt Injection] (LOW): The skill is designed to process untrusted data from an external CSV file, creating a surface for indirect prompt injection.
Ingestion points: papers/extraction_table.csv (referenced in the Workflow section of SKILL.md).
Boundary markers: Absent. The instructions do not define delimiters or provide specific warnings for the agent to ignore instructions embedded within the CSV fields.
Capability inventory: None. The skill contains only markdown instructions and no executable scripts or command-line operations.
Sanitization: Absent. There are no instructions for validating or sanitizing the content of the CSV before processing.
[No Code] (SAFE): The skill contains no executable scripts (e.g., Python, Bash, Node.js), which significantly reduces the risk of malicious payload execution or system compromise.

bias-assessor