The Agent Skills Directory

[SAFE]: No malicious patterns, obfuscation, or unauthorized data access were identified. The skill accurately documents the NIP-90 specification.
[PROMPT_INJECTION]: The skill documents a protocol surface susceptible to indirect prompt injection. Ingestion points: Untrusted data enters via kind:5000-5999 events and external URLs provided in the 'i' tag within SKILL.md and references/dvm-kinds.md. Boundary markers: The provided implementation examples do not include delimiters or specific instructions to ignore embedded prompts. Capability inventory: The service provider implementation in SKILL.md includes network fetching (fetch) and text processing (summarize). Sanitization: There is no evidence of input validation or content sanitization in the documentation or code snippets. This surface is part of the protocol's intended design and does not represent a malicious implementation by the author.

nostr-dvms