The Agent Skills Directory

[SAFE]: The skill is primarily documentation-based, providing architectural patterns and examples for agent development. All content was analyzed for malicious instructions and none were found.
[COMMAND_EXECUTION]: The provided script scripts/validate-agent.sh is a local utility for validating the structure of agent markdown files. It uses standard POSIX tools (grep, sed, awk) to perform static analysis on file frontmatter and content. It does not perform network requests or execute any content from the analyzed files.
[PROMPT_INJECTION]: The skill provides templates for creating agents that incorporate user-defined descriptions. This is a standard architectural pattern for this tool's purpose. The templates use quoting as a basic boundary marker for interpolated content.
Ingestion points: SKILL.md and examples/agent-creation-prompt.md (via the [YOUR DESCRIPTION] placeholder).
Boundary markers: The templates encapsulate user input within double quotes.
Capability inventory: The skill guides users on how to provision agents with tools such as Read, Write, Grep, and Bash.
Sanitization: No explicit sanitization logic is provided in the templates, as they are intended for use by developers during the agent creation process.

agent-development