The Agent Skills Directory

[SAFE]: The skill's description and instructions align with its stated purpose of assisting with nonfiction research, citation management, and fact-checking. Requested permissions for network and file system access are appropriate for gathering academic sources and maintaining a research database. No malicious obfuscation, exfiltration logic, or unauthorized access patterns were found. Note: The author is listed as AuthorClaw, which differs from the provided context author Ckokoski, though this does not appear to be malicious.
[NO_CODE]: The skill consists entirely of natural language instructions and YAML metadata within the SKILL.md file, with no associated scripts, binaries, or executable dependencies.
[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it is designed to ingest and process untrusted external data from academic sources and manuscripts. 1. Ingestion points: Manuscript files and external academic sources gathered via network:http. 2. Boundary markers: No explicit delimiters or instructions to ignore embedded commands are present in the process description. 3. Capability inventory: network:http, file:read, and file:write. 4. Sanitization: No sanitization or validation steps for external content are specified.

nonfiction-research