The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted external data as part of its primary function to draft proposals. While this represents a theoretical attack surface, the risk is mitigated by the skill's restricted capabilities.
Ingestion points: The skill accepts external data such as website briefs, RFPs, target audience descriptions, and SEO evidence as defined in the 'Inputs' table of SKILL.md.
Boundary markers: No explicit boundary markers or delimiters are defined in the instructions to separate untrusted input from the agent's internal logic.
Capability inventory: The skill is strictly limited by a 'Capability Contract' in SKILL.md to read-only analysis and critique. It contains no tools or scripts for network access, file system modification, or command execution.
Sanitization: There are no explicit instructions for sanitizing or escaping the content provided in the inputs.
[SAFE]: The skill implements safety best practices by including a 'Capability Contract' and 'Degraded Mode' section. These explicitly instruct the agent to remain in a read-only state, avoid making statutory or commercial commitments without authority, and handle missing information by returning qualified outlines rather than inventing facts.

website-design-proposal-strategy