swift-executor
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill instructions provide a productivity framework focused on rapid execution and unblocking progress. The content is transparent and does not contain hidden commands, obfuscation, or malicious instructions.\n- [INDIRECT_PROMPT_INJECTION]: The skill's emphasis on speed and "overcoming blockers" creates a potential vulnerability surface where the agent might prioritize execution over thorough vetting of input data.\n
- Ingestion points: The skill processes project tasks and code implementation requests within the agent's context.\n
- Boundary markers: The instructions do not specify custom delimiters for untrusted content, relying on the agent's default processing.\n
- Capability inventory: Access to tools like
Bash,Write, andEditis specified in the configuration to support rapid implementation.\n - Sanitization: The instructions include a "Decision Tree" that explicitly checks for security risks, data loss, and user harm, providing a form of behavioral sanitization during the decision-making process.
Audit Metadata