The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted external data (web app codebases, configuration files, and dependencies) to produce native architecture designs. This creates a potential surface for indirect prompt injection if malicious instructions are embedded in the analyzed code.
Ingestion points: SKILL.md (Reads package.json, routing config, state stores, API client, storage usage, and auth flows).
Boundary markers: Absent; the skill does not explicitly instruct the agent to use delimiters or ignore instructions embedded in the analyzed data.
Capability inventory: The agent generates specifications for downstream agents (Native, Scaffold, Gateway, Schema, Builder) to implement code, database schemas, and API contracts.
Sanitization: No explicit sanitization or validation of the ingested codebase data is described.
[EXTERNAL_DOWNLOADS]: The documentation contains numerous references to official technical documentation and industry-standard resources (e.g., Apple Developer, Android Developers, W3C, Material.io, and well-known developer blogs). These are categorized as safe, well-known services and do not involve executable code downloads or untrusted sources.

port