business-agent-os

Pass

Audited by Gen Agent Trust Hub on Nov 8, 2026

Risk Level: SAFEPERSISTENCEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PERSISTENCE]: The skill guides the user to implement a persistence mechanism by adding a command to their shell configuration file. It provides instructions and a shell script (go.zsh) for the user to manually add a source line to their ~/.zshrc, enabling a persistent 'go' menu for managing agent sessions across terminal reloads.
  • [INDIRECT_PROMPT_INJECTION]: The skill architecture ingests external data into the agent's context, creating a surface for indirect prompt injection.
  • Ingestion points: The skill reads files from notes/sources/ (raw articles, transcripts) and INBOX.md (raw user notes) during the /pickup, /wrap, and /review workflows.
  • Boundary markers: No explicit delimiters or instructions to ignore embedded commands are present in the command templates or file structures.
  • Capability inventory: The agent has the capability to write to the local filesystem and execute version control commands via git.
  • Sanitization: The skill does not implement sanitization or filtering for the external content before it is processed by the agent.
  • [COMMAND_EXECUTION]: The skill uses standard system utilities including git for local version control and tmux for session multiplexing. It generates a local Zsh script to automate the creation and switching of agent sessions within tmux.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 8, 2026, 12:37 AM