long-task-coordinator
Pass
Audited by Gen Agent Trust Hub on Mar 13, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by processing externalized state data.
- Ingestion points: The skill reads state files (e.g.,
docs/<topic>-state.md) to recover task status at the start of each coordination cycle, as described inSKILL.mdandreferences/workflow.md. - Boundary markers: There are no explicit delimiters or instructions provided to the agent to treat the contents of the state file as untrusted data or to ignore instructions embedded within those files.
- Capability inventory: The skill allows the use of the
Bashtool and file-system modification tools (Write,Edit) to advance tasks, which creates a risk if the agent's actions are influenced by malicious content within a state file. - Sanitization: No sanitization or validation logic for the state file content is specified before the agent processes it for decision-making.
Audit Metadata