The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill provides logic for the agent to apply when reviewing or writing Go code. This represents an attack surface where maliciously crafted Go files could attempt to influence the agent's behavior during a code review process. However, the instructions themselves promote defensive programming practices (e.g., using errors.Is instead of direct comparison) which mitigate common logic errors.
Ingestion points: The skill targets and processes all Go source files (**/*.go).
Boundary markers: No explicit delimiters or instructions to ignore embedded code comments/instructions are present.
Capability inventory: The skill allows file system operations (Read, Edit, Write, Glob, Grep) and restricted command execution (Bash for go, golangci-lint, and git).
Sanitization: No explicit sanitization of input code content is performed.
[SAFE]: The skill follows the principle of least privilege by explicitly restricting the Bash tool to a specific whitelist of necessary commands (go, golangci-lint, git) rather than allowing unrestricted shell access.
[SAFE]: External links and metadata resources point to the author's official GitHub repository (github.com/fabianoflorentino/golang-agent-skills), which is consistent with the skill's identity.

golang-pitfalls-error-handling