Create delivery resources
Draft and update pipelines, services, environments, infrastructure definitions, connectors, templates, variables, inputs, and triggers.
Codex Skills for Harness help Codex create and review Harness pipelines, services, environments, connectors, templates, policies, feature flags, cloud-cost workflows, and execution diagnostics. Download the skills for natural-language work with the Harness platform through approved repositories, APIs, and MCP tools.
task: complete harness software delivery task
inspect:
- requirements and context
- existing standards
- failure and edge cases
verify: outputs + checks + handoffHarness resources are connected across account, organization, and project scopes. Reliable work depends on current schemas, permissions, identifiers, variables, connectors, and platform modules.
These skills guide Codex to discover the live resource shape through approved Harness tooling, produce reviewable YAML or API changes, and protect credentials and production stages.
Draft and update pipelines, services, environments, infrastructure definitions, connectors, templates, variables, inputs, and triggers.
Inspect stages, steps, logs, expressions, service instances, manifests, delegates, connectors, policies, and failure strategies.
Build templates, policies, approvals, overrides, access boundaries, naming standards, and controlled promotion paths.
Summarize executions, deployment health, feature-flag usage, cloud costs, anomalies, and platform data available through authorized tools.
The steps keep context, implementation, and verification visible so the result can be reviewed and repeated.
Identify the Harness account, organization, project, module, environment, resource identifiers, permissions, delegates, and approved connectors.
Use supplied repository files or Harness describe and list operations instead of assuming fields from memory.
Create or edit the requested YAML or resource, preserve expressions and scope, protect secrets, and show the intended effect.
Run available validation, inspect resolved inputs and policy checks, require approval for controlled stages, and define rollback or failure handling.
The workflow adjusts to the project, audience, tools, and risk while preserving the same quality standard.
Build, test, scan, package, deploy, verify, approve, and promote software through reusable pipeline stages.
Manage services, environments, infrastructure, connectors, delegates, secrets references, templates, and inputs.
Trace failed steps, logs, expressions, manifests, connector access, delegates, timeouts, policy failures, and rollback.
Apply policies, approvals, access rules, templates, feature controls, cost views, and delivery reporting.
Start with one defined outcome and provide the source material, constraints, and checks that matter.
These skills are designed for people who need dependable harness software delivery work with a visible process.
Standardize resources, templates, connectors, policies, access, and shared delivery practices.
Create and debug pipelines, deployments, triggers, variables, manifests, and verification steps.
Use approved delivery paths and understand the inputs, environments, and evidence associated with a release.
Work with available cost, policy, approval, feature, and deployment data through controlled tools.
The skills require authorized Harness access for live actions. API keys, secrets, connectors, production executions, approvals, policies, feature changes, and cost decisions remain controlled by the user's permissions and review process.
Install the complete skill folder and add the project-specific context before beginning.
Keep SKILL.md with pipeline, resource, debugging, governance, and reporting workflows.
Use project scope for repository and Harness identifiers, or personal scope for reusable platform practices.
Configure the supported Harness MCP server or API access with least-privilege credentials outside the skill files.
Provide account, organization, project, modules, resource identifiers, delegates, environments, policies, and protected stages.
Practical answers about capabilities, limits, setup, and review.
Yes. They can draft and modify pipeline resources while following current schemas discovered from the repository or Harness tooling.
Live discovery and actions require an approved MCP or API connection. The skills can still review supplied YAML without one.
Yes, when logs and resource context are available through authorized access or supplied artifacts.
They are designed to reference protected credentials rather than place secret values in skill files, prompts, logs, or committed YAML.
Only with explicit authorization, suitable permissions, required approvals, validation, monitoring, and a defined failure or rollback strategy.
Clear context. Purposeful work. Relevant checks. A result others can understand.