The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted metadata from various project files to derive configuration values for the generated manifest, which can influence subsequent build and deployment commands.
Ingestion points: Extracts information from package.json, go.mod, docker-compose.yml, and other language-specific manifests (Section 2.1).
Boundary markers: Implements a refinement phase (Section 4.2) where the user reviews the draft. In autonomous mode, it creates a pull request for human verification (Section 6.2).
Capability inventory: The generated configuration is used by the okteto build, okteto deploy, and Git/GitHub CLI tools.
Sanitization: Relies on the AI agent's mapping logic and schema validation via okteto validate before execution.
[COMMAND_EXECUTION]: Employs standard developer CLI tools to validate the environment and manage the deployment lifecycle.
Evidence: Uses ls for file existence checks, okteto CLI for building and deploying, and git/gh for branch management and PR creation.
[EXTERNAL_DOWNLOADS]: Interacts with remote container registries and version control systems as part of its core developer-centric functionality.
Evidence: Standard operations include pushing/pulling images via okteto build and interacting with GitHub through the gh CLI for PR submission.

okteto-onboarding