agent-board-cli

Pass

Audited by Gen Agent Trust Hub on Jun 23, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill provides installation instructions that download binaries from the official GitHub releases of the 'stakpak' organization (github.com/stakpak/agent-board). This is consistent with the skill's authorship and intended purpose.
  • [COMMAND_EXECUTION]: The skill uses a dedicated CLI tool, agent-board, to manage task boards, cards, and comments. These commands are limited to local task management operations and do not show signs of arbitrary command execution or privilege escalation.
  • [DATA_EXPOSURE]: Application data is stored in a local SQLite database at ~/.agent-board/data.db. This is a standard and safe practice for a local-first task management application.
  • [INDIRECT_PROMPT_INJECTION]:
  • Ingestion points: The agent retrieves user-generated or agent-generated content (task descriptions, comments) from the database via agent-board get and agent-board list (SKILL.md).
  • Boundary markers: The documentation does not specify the use of delimiters when the CLI tool returns data to the agent.
  • Capability inventory: The skill can execute the agent-board CLI to modify tasks and agents on the system (SKILL.md).
  • Sanitization: No explicit sanitization of task content is mentioned, which is typical for a internal task tracking tool.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 23, 2026, 07:41 AM
Security Audit — agent-trust-hub — agent-board-cli