The Agent Skills Directory

[SAFE]: The skill files contain no executable code, remote scripts, or commands that interact with the local filesystem or sensitive environment variables. The logic is entirely prompt-based and uses platform-provided capabilities.- [PROMPT_INJECTION]: The skill has a surface for indirect prompt injection due to its core function of researching external websites. 1. Ingestion points: competitor websites, pricing pages, and review platforms (G2, Capterra) via web_search. 2. Boundary markers: none are specified to delimit external data within the synthesis prompt. 3. Capability inventory: web_search is used to aggregate data for structured battlecard generation. 4. Sanitization: no specific filtering of external content is mentioned. This risk is assessed as safe because it is inherent to the primary research purpose and the output is text for human review rather than executable code.

Competitor Intel