The Agent Skills Directory

[SAFE]: The skill performs legitimate project initialization tasks by reading standard build configuration files (pom.xml, package.json) and creating markdown documentation for agent guidance.
[SAFE]: All external resource links point to official Adobe documentation (experienceleague.adobe.com), trusted community resources (wcm.io, apache.org), or well-known package registries (npmjs.com).
[INDIRECT_PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it ingests project metadata (e.g., project name from pom.xml) and interpolates it into the AGENTS.md file which serves as instructions for future agent tasks.
Ingestion points: Root pom.xml and ui.frontend/package.json.
Boundary markers: No explicit delimiters or instructions are used to separate user-provided project strings from the agent instructions in the generated AGENTS.md.
Capability inventory: The skill performs file system discovery and file creation (AGENTS.md, CLAUDE.md).
Sanitization: The skill performs basic string manipulation (humanizing artifact IDs) but does not validate or sanitize extracted content against embedded instructions.

ensure-agents-md