galyarder-cfo-coo
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The Identity Contract section contains instructions that override the host platform's identity (e.g., Copilot or Gemini), directing the agent to maintain the Galyarder CFO/COO persona and ignore its platform-level identification.\n- [COMMAND_EXECUTION]: The skill implements an ExecutionProxy Interface that allows the agent to execute shell commands using the 'rtk' prefix (e.g., 'rtk npm test') for testing and operational optimization.\n- [INDIRECT_PROMPT_INJECTION]: The skill features a vulnerability surface where it ingests and processes untrusted external data, although it includes mandatory defensive protocols.\n
- Ingestion points: External data and web content processed via BrowserOS or third-party APIs.\n
- Boundary markers: Requires architectural data flow audits and documented legal clearance before processing new data streams.\n
- Capability inventory: Shell command execution (ExecutionProxy) and file system writes (MemoryStore/Obsidian sync).\n
- Sanitization: Explicit requirement to redact secrets and personally identifiable information (PII) before sharing context with subagents or external tools.
Audit Metadata