The Agent Skills Directory

[PROMPT_INJECTION]: Indirect Prompt Injection Surface. The skill retrieves untrusted data from external sources, including competitor websites and review platforms, using scraping and search tools.
Ingestion points: Data is ingested via firecrawl_scrape, firecrawl_search, and various DataForSEO calls as detailed in references/competitor-profiling-guidelines.md and references/competitor-profiling-tool-reference.md.
Boundary markers: The instructions do not define boundary markers (such as triple backticks or XML tags) or specific warnings to ignore instructions embedded in the external content during processing.
Capability inventory: The skill possesses the capability to write research data to the project's local file system (competitor-profiles/) and make broad network requests through configured MCP tools.
Sanitization: No explicit sanitization, validation, or filtering of the retrieved content is described before the data is synthesized into final documents.
[COMMAND_EXECUTION]: Local File System Manipulation. The skill instructs the agent to create and manage a directory structure (competitor-profiles/raw/) and write multiple markdown and JSON files to disk. While this supports the research workflow, it grants the agent ongoing write access to the local project directory.

competitors