atlas-mysql
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructs the agent to download the core
atlasskill and its associated reference documentation from the vendor's official domain atatlasgo.io. These resources are provided by the official maintainers of the tool. - [COMMAND_EXECUTION]: The instructions involve running the
atlasCLI tool for database operations including schema synchronization (atlas schema apply), migration management (atlas migrate diff), and environment validation. It also utilizesdockercommands to spin up isolated development containers for testing database changes. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to process and act upon database schema files and migration scripts which serve as untrusted ingestion points.
- Ingestion points: The skill reads HCL schema files (
*.my.hcl,*.ma.hcl) and SQL migration scripts provided by the user or project environment. - Boundary markers: The skill relies on structured HCL blocks and standard SQL syntax to delimit content, though it does not explicitly instruct the agent to ignore prompt-like instructions embedded in comments.
- Capability inventory: The agent has the capability to read files, write to the database via the CLI, and manage local Docker containers.
- Sanitization: All processed content is passed through the
atlasCLI's own internal parsers, validators, and linting engines before execution.
Audit Metadata