The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes instructions in rules/market-competitive.md to use the GitHub CLI (gh) for gathering market data, such as repository stars and release counts, which is a legitimate use of tool capabilities for market research.
[REMOTE_CODE_EXECUTION]: Python code snippets for computing financial metrics like ROI, NPV, and IRR are provided in rules/business-roi.md and rules/strategy-go-no-go.md as utility functions for the agent to perform data-driven analysis.
[PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it processes user-provided product data and maintains execution capabilities. Evidence chain:
Ingestion points: User-provided feature lists and business case queries entering the agent's context during prioritization or market analysis tasks.
Boundary markers: The skill does not explicitly define delimiters for untrusted user inputs within its templates.
Capability inventory: The skill utilizes command-line execution (GitHub CLI) and dynamic code execution (Python for mathematical modeling).
Sanitization: The skill relies on the agent's native safety guardrails and the structured nature of the provided templates for processing inputs.

product-frameworks