The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill fetches documentation and resource templates from remote vendor-controlled locations, creating an attack surface where malicious content at the source could influence the agent's actions during manifestation or documentation rendering.\n
Ingestion points: Scripts scripts/fetch-page.sh and scripts/extract-resources.sh retrieve content from openchoreo.dev and GitHub repositories.\n
Capability inventory: The skill utilizes bash, kubectl, occ, and git to manage infrastructure and source code.\n
Sanitization: Manifests are processed via awk and yq to filter specific resource components, but the template content remains largely untrusted until applied.\n
Boundary markers: The instructions do not define explicit delimiters to isolate remote content from system prompts.\n- [EXTERNAL_DOWNLOADS]: The skill is configured to download necessary resources from external domains during the setup process.\n
Fetches documentation from https://openchoreo.dev.\n
Retrieves sample resource manifests from the openchoreo/openchoreo and openchoreo/sample-gitops GitHub organizations.\n
Points to https://fluxcd.io/install.sh for the Flux CD command-line tool installation.\n- [COMMAND_EXECUTION]: The skill performs its operations by executing a variety of shell-based commands to manage cluster and repository state.\n
Executes kubectl and occ for Kubernetes cluster resource management.\n
Uses git and gh for managing Git repositories and creating pull requests.\n
Runs curl to fetch the external resources described in the downloads section.

openchoreo-platform-engineer-gitops