dagster-expert
Pass
Audited by Gen Agent Trust Hub on May 18, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill provides detailed documentation for using the
dgCLI to manage Dagster projects, including scaffolding definitions, materializing assets locally withdg launch, and interacting with the Dagster Plus API for deployment management. - [EXTERNAL_DOWNLOADS]: The instructions guide the installation of official Dagster integration libraries (e.g.,
dagster-dbt,dagster-aws,dagster-fivetran) and the use ofcreate-dagsterfor project scaffolding. These resources are hosted on official registries (PyPI) and originate from the trusted vendor. - [CREDENTIALS_UNSAFE]: The documentation describes standard procedures for managing secrets and environment variables using the CLI (e.g.,
dg api secret get,dg plus pull env). It includes best practices such as using.envfiles and ensuring they are added to.gitignoreto prevent credential exposure. - [INDIRECT_PROMPT_INJECTION]: The skill instructions mandate that the agent read local project files and reference documentation before providing answers. This represents a standard surface for indirect prompt injection if a processed codebase were to contain malicious instructions designed to influence the agent's output. The risk is mitigated by the expert-focused, documentation-centric nature of the skill.
Audit Metadata