gcp-cloud-build
Pass
Audited by Gen Agent Trust Hub on Oct 2, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes shell commands using the gcloud CLI and jq to interact with Google Cloud Platform services. These are intended administrative actions for resource discovery and management.
- [INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection through external build logs. * Ingestion points: The get_build_logs helper function in SKILL.md fetches the last 100 lines of build logs into the agent context. * Boundary markers: Absent; there are no explicit instructions or delimiters defined to prevent the agent from following instructions embedded within the logs. * Capability inventory: The agent has access to gcloud commands to manage triggers, builds, and worker pools. * Sanitization: Metadata is filtered using jq, but build logs are retrieved as raw text without sanitization for potentially malicious instructions.
Audit Metadata