The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a bundled local script, scripts/gerar-swot-visual.py, to convert analysis data into an interactive HTML visualization. This script uses standard Python libraries to parse markdown and generate a local report, which is a legitimate and restricted functional requirement.
[INDIRECT_PROMPT_INJECTION]: The skill features a surface for indirect prompt injection due to its core functionality of processing external data. 1. Ingestion points: External information enters the system through market research performed by agente-pesquisa.md and project descriptions provided by the user. 2. Boundary markers: The instructions do not employ specific markers to delimit external content within prompts. 3. Capability inventory: The skill can perform web searches and execute its local visualization script. 4. Sanitization: The multi-agent debate and synthesis process by the agente-compilador.md provide an inherent review layer, and the Python script uses json.dumps to structurally validate data injection into the HTML template.
[EXTERNAL_DOWNLOADS]: The HTML visualization template utilized by the skill and generated by its Python script references CSS resources from the official Google Fonts CDN.

analise-swot