flow-next

Warn

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION] (MEDIUM): The skill executes a bundled script flowctl located in the plugin root. It uses shell-based data passing methods like heredocs (<<'EOF') and temporary files in /tmp/. If input titles or descriptions are not properly sanitized before being passed to these shell commands, it could lead to command injection.
  • [PROMPT_INJECTION] (MEDIUM): The skill is vulnerable to Indirect Prompt Injection (Category 8). It ingests untrusted data from the .flow/ directory using commands like $FLOWCTL cat and $FLOWCTL show.
  • Ingestion points: Reads epic and task specifications from the filesystem.
  • Boundary markers: None. There are no delimiters or instructions to ignore embedded commands in the processed data.
  • Capability inventory: Has the ability to execute shell commands and modify local files via $FLOWCTL.
  • Sanitization: None detected. Malicious instructions embedded in a task's markdown description could influence the agent's next actions.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 16, 2026, 04:27 AM