10-andruia-skill-smith
Pass
Audited by Gen Agent Trust Hub on Sep 6, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to perform local file system manipulations, which include directory creation and writing contents to new files.
- Evidence: 'Crear la carpeta física en D:...\agentic-awesome-skills\skills' in SKILL.md.
- Evidence: 'Escribir el archivo README.md en dicha carpeta' in SKILL.md.
- [INDIRECT_PROMPT_INJECTION]: The skill generates new sets of instructions based on external user input, creating a surface where malicious prompts could be embedded into new skill files.
- Ingestion points: Phase 1 (ADN de la Skill) in SKILL.md collects 'Nombre Técnico', 'Rol Experto', and 'Outputs Clave' from user input.
- Boundary markers: No specific delimiters or warnings to ignore embedded instructions are present for the generated content.
- Capability inventory: The skill possesses file system write capabilities and folder creation as defined in Phase 3 of SKILL.md.
- Sanitization: No validation or sanitization logic is present to filter user-provided roles or output requirements before they are materialized into skill files.
Audit Metadata