The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill contains an inherent attack surface because its primary purpose is to review and audit Go source code, which may contain malicious instructions or edge cases designed to exploit the analysis process.
Ingestion points: The skill ingests untrusted data from the local filesystem using Read, Glob, and Grep tools to find and examine .go files.
Boundary markers: There are no specific instructions or delimiters provided to the agent to treat the analyzed code as untrusted data or to ignore any embedded natural language instructions within that code.
Capability inventory: The skill utilizes shell execution capabilities via the Bash tool to run commands such as go test -race, go vet, golangci-lint, and gosec. While restricted to specific binaries, executing tests on untrusted code can trigger code execution (e.g., via malicious test cases).
Sanitization: The skill does not define any sanitization or validation logic for the content of the files before they are processed by the agent or the underlying shell tools.

go-defensive-coding