business-operations-skills
Pass
Audited by Gen Agent Trust Hub on Oct 31, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data which creates a potential surface for indirect prompt injection.
- Ingestion points: The orchestrator and its sub-skills are designed to read and analyze local business data, including vendor catalogs, process transcripts, SOP source documents, and financial exports (CSV/JSON) found in the user's working directory.
- Boundary markers: The skill does not define specific prompt delimiters for external data, but uses a 'forked context' strategy to isolate heavy data ingestion from the primary conversation thread.
- Capability inventory: The skill possesses the capability to execute local Python scripts and write analysis artifacts (Markdown, CSV, JSON) to the user's working directory.
- Sanitization: There is no explicit mention of data sanitization or escaping mechanisms for the ingested business documents.
- [COMMAND_EXECUTION]: The skill facilitates the execution of local Python tools for operational analysis (e.g., capacity planning, spend audits). The implementation is restricted to the Python standard library (stdlib-only) and explicitly prohibits external API calls, significantly reducing the attack surface.
- [DYNAMIC_EXECUTION]: Analysis tasks are performed via runtime execution of localized scripts to generate artifacts and digests. These operations are intended for deterministic data processing rather than arbitrary code execution.
Audit Metadata