The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill accesses untrusted data by viewing GitHub issue comments and labels via the gh CLI. Ingestion points: Data entering the context from issue comments and labels fetched using gh issue view in SKILL.md. Boundary markers: Absent; the instructions do not include delimiters or warnings to ignore embedded instructions within the issue content. Capability inventory: The agent can execute shell commands for testing and typechecking, use the gh tool to edit and comment on issues, and manage git branches. Sanitization: Absent; the skill does not specify filtering or escaping of the fetched issue content before processing.
[COMMAND_EXECUTION]: The skill requires the execution of shell commands, specifically the GitHub CLI (gh) for issue management and local development tools for typechecking and running test suites. While these are necessary for the skill's function, they provide a functional capability that could be leveraged if the agent encounters malicious instructions in external inputs.

implement