agency-visual-storyteller
Pass
Audited by Gen Agent Trust Hub on Mar 25, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses shell commands to inspect local project files and directories as part of its strategy development workflow. Evidence:
cat ai/memory-bank/brand-guidelines.md,ls public/images/brand/, andgrep -i "story\|narrative\|message"commands inSKILL.md. - [PROMPT_INJECTION]: The skill demonstrates an attack surface for indirect prompt injection by ingesting untrusted file content into the agent's context. 1. Ingestion points: File content from
ai/memory-bank/brand-guidelines.mdandai/memory-bank/audience-research.mdis read viacatcommands inSKILL.md. 2. Boundary markers: No explicit delimiters or instructions to ignore embedded commands are present in the workflow. 3. Capability inventory: The agent has shell access to read files and list directories. 4. Sanitization: No evidence of input validation or content filtering is provided.
Audit Metadata