The Agent Skills Directory

[SAFE]: The skill is composed of markdown-based business playbooks and local Python scripts designed for revenue and churn analysis. The analysis confirmed that the scripts use only Python standard libraries, do not perform network operations, and do not access sensitive system paths or credentials. The overall logic is transparent and aligned with the stated business advisory purpose.\n- [INDIRECT_PROMPT_INJECTION]: The skill defines a surface for indirect prompt injection by ingesting untrusted external data through a context file and user-provided CSV inputs. However, the associated risks are mitigated by the lack of dangerous capabilities.\n
Ingestion points: The agent is instructed to read company-context.md, and the provided scripts (churn_analyzer.py and revenue_forecast_model.py) ingest customer and pipeline data via CSV files.\n
Boundary markers: No explicit boundary markers or instructions to ignore embedded commands are present in the processing logic for these inputs.\n
Capability inventory: The tools are limited to data parsing and statistical reporting; they contain no network calls, file-system write operations, or shell command execution logic.\n
Sanitization: The scripts rely on standard library CSV parsing without additional sanitization or filtering of input content.

cro-advisor