sop-creator
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill incorporates external content from user descriptions and local files into its processing logic, which creates a potential surface for indirect prompt injection.
- Ingestion points: The skill ingests untrusted data from the $ARGUMENTS variable, subsequent user messages, and the FOUNDER_CONTEXT.md file (referenced in SKILL.md).
- Boundary markers: No explicit delimiters, wrappers, or 'ignore embedded instructions' warnings are defined to separate the skill's instructions from the content of the ingested data.
- Capability inventory: The skill is limited to text generation and the AskUserQuestion tool; it lacks capabilities for network access, file system writes, or shell command execution across all scripts (SKILL.md).
- Sanitization: The instructions do not specify any validation, escaping, or filtering mechanisms for the ingested process descriptions or business context files.
Audit Metadata