The Agent Skills Directory

[SAFE]: The skill contains no malicious instructions, obfuscation, or unauthorized commands. Its functionality is limited to searching the web and reading local files as defined in the allowed-tools configuration.
[SAFE]: A vulnerability surface for indirect prompt injection exists because the skill processes external content from web searches. However, this is the primary purpose of the skill and no specific exploitation patterns were found. Ingestion points: Output from mcp__perplexity-ask__perplexity_ask. Boundary markers: Absent. Capability inventory: Read, Grep, Glob. Sanitization: Absent.

researching-web