flyio-core-workflow-b
Pass
Audited by Gen Agent Trust Hub on Sep 13, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns, prompt injections, or data exfiltration attempts were detected. The skill provides standard operational procedures for the Fly.io CLI.
- [FRONTMATTER]: The skill implements security best practices by using the
allowed-toolsfield to limit the agent's execution environment to specific, relevant command-line tools (flyandpsql). - [CREDENTIALS_SAFE]: The skill explicitly warns the agent not to include connection strings, records, or keys in its output. It uses standard environment variable interpolation (e.g.,
"$DATABASE_URL") for database connections, which avoids hardcoding sensitive information.
Audit Metadata