The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill contains instructions for the agent to read, audit, and extend existing command structures and source code (defined in the "Extend" and "Review" modes). This represents a vulnerability surface where the agent could ingest adversarial instructions hidden within the codebase it is analyzing.
Ingestion points: The agent is instructed to read the project's "current command tree" and source files from the local filesystem during build and review tasks (SKILL.md).
Boundary markers: The instructions do not include specific delimiters or warnings to help the agent distinguish between legitimate source code and potential embedded instructions within the files being processed.
Capability inventory: The skill enables the agent to generate and modify Go code capable of file system access, network service management, and process control (SKILL.md).
Sanitization: The skill lacks specific sanitization for the agent's ingestion process, though it correctly advocates for the use of Cobra's built-in argument and flag validation for the tools being developed.
[EXTERNAL_DOWNLOADS]: The skill references several industry-standard Go libraries and tools, including github.com/spf13/cobra, github.com/spf13/viper, and github.com/charmbracelet/bubbletea. These are well-known, widely used packages in the Go community for building command-line interfaces and managing configuration.

golang-cli