The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a local Python script scripts/gerar-swot-visual.py to generate an HTML visualization of the analysis results. This is a standard functional component for reporting.
[INDIRECT_PROMPT_INJECTION]: The skill has an ingestion surface as it processes user-provided project descriptions and answers to five context-gathering questions. It includes specific sub-agents to debate and refine this content, which serves as a natural review layer against malicious instructions embedded in user data. The capability to write files is restricted to a structured markdown template and a generated HTML file.

analise-swot