The Agent Skills Directory

[SAFE]: The skill is a legitimate automation tool for generating Harness pipeline configurations using provided schemas and best practices.
[SAFE]: Instructions include clear security guidelines, such as the requirement for approval stages with 'disallowPipelineExecutor' enabled to prevent unauthorized self-approval of deployments.
[SAFE]: The skill explicitly forbids hardcoding sensitive information like AWS account IDs, regions, or cluster identifiers, requiring the agent to ask the user for these values instead of using placeholders.
[SAFE]: Promotes the use of Harness native steps which provide built-in error handling and reduced shell injection risk compared to raw scripts.
[DATA_EXPOSURE]: The skill analyzes local project manifest files (e.g., package.json, go.mod, pom.xml) to auto-detect configuration requirements. This data access is limited to the project's own files for the intended purpose of CI/CD setup and does not target sensitive system or user directories.
[INDIRECT_PROMPT_INJECTION]: The skill's codebase analysis involves reading untrusted files from a repository. While this creates an attack surface for indirect prompt injection, the risk is mitigated by a mandatory instruction to confirm all auto-detected settings with the user before generating or applying any configurations.

create-pipeline