The Agent Skills Directory

[PROMPT_INJECTION]: The skill retrieves and processes external data from the Promptwatch API, which presents a surface for indirect prompt injection.
Ingestion points: Data is fetched using the list_content_gap_prompts and get_content_gap_latest tools in SKILL.md.
Boundary markers: The instructions lack delimiters or markers to separate untrusted external content from the agent's logic.
Capability inventory: The skill utilizes the create_content tool, which can perform actions based on the processed data.
Sanitization: No explicit sanitization or validation of the data retrieved from the external source is mentioned.
[EXTERNAL_DOWNLOADS]: The skill documentation identifies the Promptwatch MCP server and API endpoints (promptwatch.com) as required resources. These are official vendor domains.
[COMMAND_EXECUTION]: The skill mentions a setup-promptwatch-mcp utility for initial configuration, which is a vendor-provided tool.
[SAFE]: The workflow includes a critical safety step requiring explicit user approval before the agent calls the create_content tool, preventing silent generation.

content-gap-audit