The Agent Skills Directory

[SAFE]: No malicious patterns, obfuscation, or sensitive data access were detected. The skill instructions are limited to defining a research review workflow.
[NO_CODE]: The skill does not include any executable scripts, shell commands, or external dependencies, which minimizes the security risk.
[PROMPT_INJECTION]: The skill handles user-provided research artifacts, representing a surface for indirect prompt injection.
Ingestion points: User-provided research papers and drafts are processed by the agents mentioned in the workflow.
Boundary markers: There are no explicit markers or instructions to ignore embedded commands within the ingested research data.
Capability inventory: The skill is restricted to generating output in the outputs/ directory. No tools for file system modification beyond the output folder, network access, or command execution are requested.
Sanitization: No input validation or sanitization procedures are specified for the external content.

research-review