The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external JSON data, such as RFP responses, customer usage metrics, and sales pipeline exports, to generate scores and recommendations that drive subsequent agent actions.
Ingestion points: Data is ingested via scripts like rfp_response_analyzer.py, health_score_calculator.py, and pipeline_analyzer.py which read from files in the assets/ directory.
Boundary markers: The provided instructions do not include explicit delimiters or safety instructions to prevent the agent from interpreting embedded text in the data files as commands.
Capability inventory: The skill can execute local Python scripts, perform file system read operations, and invoke system utilities like Pandoc for document conversion.
Sanitization: No specific sanitization or strict schema validation mechanisms are described for the external JSON inputs.
[COMMAND_EXECUTION]: The skill utilizes several local Python scripts to automate business analysis. It also employs inline Python execution (python -c) to evaluate script outputs and control the workflow, such as determining if the agent should proceed to the next phase of a project based on calculated scores.
[EXTERNAL_DOWNLOADS]: The skill includes instructions to install the pandoc document converter using well-known system package managers like Homebrew (brew) or APT (apt). These represent downloads from established, trusted software repositories.

business-growth-skills