manage-project

Pass

Audited by Gen Agent Trust Hub on Sep 2, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses the andrue-cli command-line tool to manage project state, initialize tasks, and serve a local operator board at http://127.0.0.1:4373. These commands are standard operations for the skill's intended purpose of task management and do not involve arbitrary shell injection or dangerous system modifications.
  • [REMOTE_CODE_EXECUTION]: The skill references an external CLI package @andrue/cli as a requirement. Per the provided instructions, this is a vendor-owned resource associated with the author and is treated as a standard dependency for the skill's operation.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes project-request and project-context from external sources. It includes strong mitigation patterns such as strict 'Ready gates', evidence-based reviews, and instructions to 'Inspect evidence rather than accepting agent claims', which significantly reduces the surface for indirect injection attacks.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 2, 2026, 06:58 PM
Security Audit — agent-trust-hub — manage-project