skills/marcus/td/td-task-management/Gen Agent Trust Hub

td-task-management

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • Indirect Prompt Injection (LOW): The skill facilitates the storage and retrieval of task-related metadata and logs which serve as a persistent vector for untrusted data to enter the agent's context window.\n
  • Ingestion points: The agent is instructed to call td usage, td context <id>, and td show <id> to retrieve work state, which reads from user-controllable files in the .todos/ directory.\n
  • Boundary markers: There are no defined delimiters or instructions to treat handoff notes or task descriptions as data rather than instructions, increasing the risk of the agent obeying embedded commands.\n
  • Capability inventory: The agent using this skill is expected to perform development tasks, meaning it likely has access to file-write and command execution capabilities that could be abused via malicious task data.\n
  • Sanitization: The documentation provides no guidance on sanitizing input logs or validating the contents of handoffs before they are processed by the next agent session.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:29 PM