The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill traces and analyzes external codebase implementation details, which creates a surface for indirect prompt injection if the analyzed source code contains malicious instructions intended to influence the agent's output. * Ingestion points: The Explorer workflow reads source files, type definitions, and configurations using the runtime's file-search and read tools. * Boundary markers: The prompt templates (explorer-prompt.md, explainer-prompt.md) lack specific delimiters or instructions to ignore instructions found within the analyzed code. * Capability inventory: The skill is limited to reading files and synthesizing explanations; it does not request network access or write permissions. * Sanitization: No explicit sanitization or filtering of the ingested source code is performed before synthesis.
[NO_CODE]: The skill consists entirely of markdown documentation and prompt templates, with no executable scripts or binaries included.

how