The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to use several CLI tools, including git, gh, glab, and jq, to explore the repository state and interact with external issue trackers. These commands are used to read configuration (git remote -v, .git/config), manage issues (create, view, list, comment), and handle labels on GitHub and GitLab.
[INDIRECT_PROMPT_INJECTION]: The skill establishes an attack surface for indirect prompt injection by design. It instructs the agent to read and process content from external sources, specifically GitHub/GitLab issue descriptions and comments. If an issue or comment contains malicious instructions, an agent consuming this data (e.g., during the triage process defined in the templates) might be influenced by that untrusted input.
Ingestion points: Commands like gh issue view <number> --comments and glab issue view <number> --comments pull untrusted data from remote issue trackers into the agent's context.
Boundary markers: The provided templates do not specify explicit delimiters or "ignore instructions" warnings for the content fetched from issues.
Capability inventory: The skill uses commands for file reading, file writing, and repository interaction (git, gh, glab, jq).
Sanitization: There is no evidence of sanitization or filtering of the fetched issue content before it is processed by the agent.

setup-matt-pocock-skills