flow-plan-review

Warn

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • COMMAND_EXECUTION (MEDIUM): The skill instructs the agent to execute shell commands using the rp-cli tool. It interpolates user-provided #$ARGUMENTS directly into the command execution strings (e.g., rp-cli -w <id> -e 'builder ...'). This pattern presents a risk of command injection if the input contains shell metacharacters or maliciously crafted strings intended to escape the intended command context.\n- PROMPT_INJECTION (LOW): The skill is susceptible to Indirect Prompt Injection (Category 8) because it ingests and processes untrusted external data in the form of plan files and architecture documents.\n
  • Ingestion points: Implementation plans (plans/<slug>.md) and architecture docs provided via arguments.\n
  • Boundary markers: Absent. The instructions do not use delimiters or provide "ignore embedded instructions" warnings when passing data to the context builder.\n
  • Capability inventory: The skill utilizes rp-cli to perform file reads, directory searches, and send messages to a separate chat environment.\n
  • Sanitization: Absent. There is no evidence of sanitization, validation, or escaping of the file content before it is interpolated into the context building and chat delegation phases.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 17, 2026, 06:22 PM