The Agent Skills Directory

[SAFE]: No security threats were identified across any of the 11 categories. The skill is composed entirely of markdown-based instructions, protocols, and templates for structured information processing.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted external data (books, excerpts, and user-provided summaries). However, it includes robust boundary markers and protocols (e.g., source-integrity-protocol.md, source-boundary-rules.md) that instruct the agent to distinguish clearly between source text and framework analysis, label source reliability levels (A, B, C, D), and avoid adopting the source's claims as authoritative facts without verification. The capability inventory is restricted to standard text generation without access to sensitive system tools or network operations.
[REMOTE_CODE_EXECUTION]: There are no patterns of remote code execution, package installations, or external script downloads.
[DATA_EXFILTRATION]: No network operations or credential harvesting patterns were found. The skill operates on local context and user-provided data.
[DYNAMIC_CONTEXT_INJECTION]: No use of dynamic execution placeholders (e.g., shell command execution at load time) was detected in the skill configuration.

crossframe-notebook