The Agent Skills Directory

[SAFE]: The skill follows standard financial analysis protocols to identify and explain budget and period-over-period variances. Its activities are confined to processing data from the internal-gl tool and producing a summary table, which is consistent with its intended purpose.
[SAFE]: There is no evidence of prompt injection, obfuscation, or privilege escalation. The skill does not utilize tools for network communication or system-level modification, ensuring that sensitive financial information remains within the agent's authorized context.
[INDIRECT_PROMPT_INJECTION]: The skill ingests data from General Ledger records via the internal-gl tool (ingestion point). While it lacks explicit boundary markers or sanitization for this data, it possesses no exploitable capabilities such as file-write or network operations (capability inventory). Consequently, the surface for indirect instruction execution is non-exploitable and safe.

variance-commentary