The Agent Skills Directory

[SAFE]: The skill performs routine document processing tasks, including scanning and updating citation formats on 'brain pages' using standard platform tools.
[SAFE]: No obfuscation, hidden URLs, or suspicious encoding were detected in the skill instructions or associated routing evaluation data.
[SAFE]: There are no external network calls, remote dependency downloads, or command executions present in the skill configuration.
[SAFE]: While the skill interacts with untrusted data (brain pages), which represents a theoretical indirect prompt injection surface, it lacks the high-risk capabilities (e.g., network exfiltration, arbitrary shell execution) necessary to facilitate a significant attack.
Ingestion points: Reads page content via get_page as defined in SKILL.md.
Boundary markers: No specific delimiters or safety instructions are defined for processing the page content.
Capability inventory: Authorized to use search, get_page, put_page, and list_pages for document management.
Sanitization: No explicit sanitization or validation of the input content is described, though the output is restricted to formatting fixes.

citation-fixer