agkan-icebox-subtask
Pass
Audited by Gen Agent Trust Hub on Apr 8, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes the 'agkan' CLI tool to interact with task data. It performs read operations (
agkan task get) and write operations (agkan task update). These commands are necessary for the skill's stated purpose of managing project tasks. - [PROMPT_INJECTION]: The skill has a surface for indirect prompt injection because it retrieves and processes task descriptions which could contain instructions intended to influence the agent's decision-making logic.
- Ingestion points: Task details (potentially including untrusted user-generated content) are ingested via the
agkan task get <id> --jsoncommand inSKILL.md. - Boundary markers: None identified; the agent is instructed to evaluate the task data directly against specified criteria.
- Capability inventory: The skill allows the agent to modify task status and content using
agkan task updateinSKILL.md. - Sanitization: No explicit sanitization or filtering of the task content is performed prior to the evaluation step.
Audit Metadata