The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes data ingested from external CLI tool outputs and project records, which represents a potential attack surface for indirect prompt injection.
Ingestion points: Data is retrieved through commands like objects, seek, query list, and records get as specified in SKILL.md.
Boundary markers: The instructions do not define explicit delimiters or instructions to ignore potential commands embedded within project or task names.
Capability inventory: The skill has functional write capabilities including project task create, project task update, and project work-hours log (as listed in SKILL.md).
Sanitization: Although the skill mandates the use of structured JSON for --data and checking --help for field definitions, it lacks specific sanitization logic for content originating from external records.

qiqi-project