The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests and analyzes untrusted data from a user's repository, which creates a surface for indirect prompt injection attacks where malicious instructions could be embedded in configuration files. * Ingestion points: The skill reads okteto.yaml, Dockerfile, and dependency manifests including package.json, go.mod, pom.xml, and requirements.txt (SKILL.md, Operating Rule 1). * Boundary markers: The skill defines strict Operating Rules and a Pre-return review checklist to provide structure and limit the agent's susceptibility to accidental obedience of embedded data. * Capability inventory: The agent is authorized to execute okteto validate, okteto build <service>, and okteto deploy --wait (SKILL.md, Operating Rule 8). * Sanitization: The skill mandates that all output manifests pass okteto validate and a ten-point review checklist.
[COMMAND_EXECUTION]: The skill utilizes the okteto CLI tool to perform infrastructure and development environment management tasks. These commands (validate, build, deploy) are essential to the skill's stated purpose of optimizing and applying Okteto manifest changes.

okteto-manifest-optimizer