landing-page-art-direction
Pass
Audited by Gen Agent Trust Hub on Sep 11, 2026
Risk Level: SAFEDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [DATA_EXFILTRATION]: The skill instructions in
references/direction-axes.mddirect the agent to interact with~/.landing-skills/design-log.json. Accessing and modifying files in the user's home directory to maintain state across projects is a sensitive file system operation that could be leveraged for data tracking or exposure. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and act upon a user-provided 'brief' which presents an attack surface for indirect prompt injection.
- Ingestion points: User-provided 'Design Read' or brief description in
SKILL.md. - Boundary markers: None. There are no instructions to the agent to treat the brief as untrusted data or to use delimiters.
- Capability inventory: The agent is empowered to write to
DESIGN.mdand read/write the~/.landing-skills/design-log.jsonfile. - Sanitization: No validation or sanitization of the brief is specified, allowing potential malicious instructions to influence the agent's file-writing actions.
Audit Metadata