The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it processes untrusted source code and repository instructions to generate documentation.
Ingestion points: The skill reads exported declarations, implementations, tests, and existing documentation from the Go codebase, along with repository-specific instructions (SKILL.md, Steps 1 and 2).
Boundary markers: The instructions do not define specific delimiters or instructions to ignore potential commands embedded within the processed source code.
Capability inventory: The agent can write to files, execute go doc to check output, run go test to verify examples, and use repository-defined formatting tools (SKILL.md, Step 7).
Sanitization: No specific sanitization or escaping of external content is mentioned before it is included in documentation strings.
[COMMAND_EXECUTION]: The skill utilizes local command-line tools to perform verification and formatting tasks.
Evidence: The skill directs the agent to execute go doc for inspection and go test for verification, along with repository-specific formatting tools (SKILL.md, Step 7; references/examples.md, 'Validation').
[DYNAMIC_EXECUTION]: The skill involves the execution of code within the package being documented through the testing framework.
Evidence: The skill requires running Example tests and package tests to verify documentation accuracy and recipe validity (SKILL.md, Step 7; references/examples.md, 'Validation').

godoc