agent-skill-generator
Pass
Audited by Gen Agent Trust Hub on Jun 23, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill documentation describes a mechanism (
generate_skill_with_user_input {ARGUMENTS}) that ingests raw user data to generate agent instructions and code. This creates a surface for indirect prompt injection where an attacker could provide crafted input to influence the behavior of the newly generated skill or the current agent's execution context. - [COMMAND_EXECUTION]: The skill uses and references the 'Bash' tool to execute operations. While these are part of the intended functionality for skill management, the interpolation of user-supplied arguments into command-line patterns without explicit sanitization instructions presents a potential command injection risk.
- [DATA_EXPOSURE_AND_EXFILTRATION]: The skill has the capability to write and edit files. While no explicit malicious exfiltration is present, the capability inventory includes 'Write' and 'Edit' tools which, when combined with 'Bash', could be used to manipulate or expose sensitive files if the agent is misled by malicious user input.
Audit Metadata