deployment-pipeline-design
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill is primarily instructional, offering design patterns and YAML templates for deployment orchestration and safety gates.
- [COMMAND_EXECUTION]: Includes standard shell command templates for tools such as kubectl, docker, and curl to demonstrate pipeline stages and rollback strategies. These commands are consistent with the skill's purpose in a DevOps context.
- [EXTERNAL_DOWNLOADS]: References the flagsmith Python library for feature flag management. This is a recognized service for the described functionality.
- [INDIRECT_PROMPT_INJECTION]: The skill documents a surface for data ingestion where automated verification logic processes external metrics.
- Ingestion points: Prometheus API query results fetched via curl in SKILL.md.
- Boundary markers: None.
- Capability inventory: curl (network), kubectl (command execution), and bc (calculation) are used to process and act on external data.
- Sanitization: None. While this constitutes a vulnerability surface, it reflects standard automation patterns for CI/CD environments.
Audit Metadata