sdp-onboard
Pass
Audited by Gen Agent Trust Hub on Aug 31, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes shell commands to manage the onboarding lifecycle. This includes cloning repositories (
git clone), modifying configuration files usingperlstring replacement, and managing Kubernetes resources viakubectl,flux, andhelm. These operations are consistent with the skill's primary purpose of application deployment and infrastructure management. - [INDIRECT_PROMPT_INJECTION]: The skill implements a workflow that pulls code scaffolding from template repositories which could theoretically contain malicious instructions.
- Ingestion points: Project templates are fetched from internal GitLab repositories (e.g.,
1cijfer-ho) into the agent's context during the onboarding process. - Boundary markers: The skill does not define specific delimiters to separate template data from the agent's internal instruction set.
- Capability inventory: The environment possesses high-capability tools including
git,kubectl, andflux, which are used to commit code and manage live Kubernetes clusters. - Sanitization: Automated string replacement (
perl -pi -e) is used for project renaming, but the skill does not include specific validation or sanitization for the logic or contents of the fetched templates.
Audit Metadata