The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes various shell scripts and system utilities, including the bundled analyze-base-images.sh and external scripts like updateBaseImages.sh and getLatestImageTags.sh to perform image analysis and updates.
[EXTERNAL_DOWNLOADS]: The skill's workflow depends on cloning external repositories from the redhat-developer organization, specifically rhdh, rhdh-operator, and rhdh-downstream, which provide the necessary build scripts and target files.
[COMMAND_EXECUTION]: The skill uses established developer tools such as skopeo, jq, gh, and git to interact with the Red Hat Container Registry and manage Git operations for pull requests.
[PROMPT_INJECTION]: The analyze-base-images.sh script contains a vulnerability surface where it parses comments from Containerfile and Dockerfile to extract configuration for the getLatestImageTags.sh script. While shell quoting is used to mitigate direct command execution, this represents an ingestion point for untrusted data from repository files.
Ingestion points: Comments in Containerfile or Dockerfile parsed via regex to extract image URLs and tag filters.
Boundary markers: None; the script relies on specific comment formatting conventions.
Capability inventory: The script performs registry queries via skopeo and executes the getLatestImageTags.sh script with extracted arguments.
Sanitization: The resolve_glit_tag function extracts content after a # delimiter but does not perform extensive validation before passing it as a command-line argument.

update-base-image