The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The file REFERENCES.md contains a link to https://verseo.framer.website/, which has been identified by automated security scanners as a phishing domain. The skill instructions in SKILL.md and demo/PROMPT.md direct the agent to analyze the behavior and quality of these external references, potentially exposing the agent to malicious content.
[METADATA_POISONING]: Automated security scans have flagged SKILL.md and REFERENCES.md with malware reputation alerts (FileRepMalware). The presence of these flags in the skill's primary documentation and instructions suggests a potential compromise or the inclusion of malicious patterns.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted external data to generate functional landing page content.
Ingestion points: The skill ingests user-provided "source notes", "interview notes", and "product briefs" as described in SKILL.md and the composer input field in demo/index.html.
Boundary markers: No specific boundary markers or instructions to ignore embedded commands are present in the skill's guidelines.
Capability inventory: The skill instructs the agent to generate and preview functional HTML, CSS, and JavaScript. The ability to generate executable frontend code based on external input creates a potential vector for cross-site scripting (XSS) if the input contains malicious payloads.
Sanitization: There are no requirements or instructions for sanitizing or validating the ingested content before it is used to generate output code.

product-proof-saas