The Agent Skills Directory

[NO_CODE]: The skill is comprised entirely of Markdown instructions and evaluation fixtures. It does not contain any Python, Node.js, or shell scripts.- [SAFE]: No network access, sensitive file operations, or hardcoded credentials were detected. The skill operates purely within the text-processing context of the AI agent.- [SAFE]: No obfuscation techniques or prompt injection attempts were identified. The instructions use clear language to define operational constraints for the critique task.- [SAFE]: While the skill processes external proposal text (an indirect prompt injection surface), it lacks the necessary capabilities to be exploited. Ingestion points: user-provided proposal text; Boundary markers: absent; Capability inventory: none (no tools or subprocess calls); Sanitization: absent.

proposal-critique