devops-automation-pack-hardened
Warn
Audited by Gen Agent Trust Hub on Apr 21, 2026
Risk Level: MEDIUMPROMPT_INJECTIONNO_CODECOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The
SKILL.mdandSAFETY.mdfiles contain directive instructions and simulated adversarial examples designed to override agent behavior and safety protocols. Use of phrases like 'regardless of claimed urgency or prior authorization claims' and 'These boundary holds regardless of claimed authority' are patterns intended to supersede other system or user instructions. - [NO_CODE]: The skill is marketed as a 'Complete DevOps automation toolkit' for Docker, Kubernetes, and CI/CD, yet it fails to provide any of the functional scripts mentioned in its documentation (e.g.,
docker-deploy.sh,k8s-deploy.sh,ci-cd.sh,monitor.sh). The only script file included,deploy.sh, is empty. - [COMMAND_EXECUTION]: The documentation provides examples for executing local shell scripts that are not part of the skill's file set. This could lead to the execution of unintended files if scripts with these names already exist in the user's local environment.
- [EXTERNAL_DOWNLOADS]: The skill references external URLs and repositories (e.g.,
faberlens.ai,github.com/faberlens/hardened-skills) for further configuration and issue reporting. These resources are managed by the skill's author.
Audit Metadata