The Agent Skills Directory

[SAFE]: The skill uses official Amplemarket MCP tools and standard web search capabilities to perform its stated functions. Its behavior is consistent with the author's provided documentation.
[SAFE]: No obfuscation, hardcoded credentials, or suspicious exfiltration patterns were detected. Data handling is limited to the scope of ICP refinement.
[PROMPT_INJECTION]: The skill ingests untrusted data from web search results and company enrichment API calls, creating a surface for indirect prompt injection. This is a low-risk concern inherent to the skill's research-based functionality.
Ingestion points: WebSearch results and mcp__claude_ai_Amplemarket__* API tool outputs (enrichment, search results).
Boundary markers: Not explicitly implemented in the prompt instructions to separate untrusted tool output from internal logic.
Capability inventory: Tool access includes web searching, firmographic enrichment, and lead list management within the Amplemarket platform.
Sanitization: The agent is instructed to analyze and extract patterns from the ingested content; no explicit sanitization for subverting instructions is noted.

icp-refiner