dramagic-character-pack
Pass
Audited by Gen Agent Trust Hub on Oct 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to generate a character manifest containing a
base_instructionfield. This field is intended to be used by an AI agent at runtime to define roleplay behavior, creating a potential surface for indirect prompt injection if a malicious user provides instructions that attempt to override the consuming agent's safety guardrails. - Ingestion points: User-provided character briefs, personality descriptions, and visual directions as defined in
SKILL.md. - Boundary markers: The skill documentation in
format.mdexplicitly categorizes tone and personality text as "untrusted roleplay data" and instructs that it should not be treated as executable instructions or permission to run tools. - Capability inventory: The skill manages file system operations, including creating working directories, writing JSON manifest files, and organizing image assets as described in
SKILL.md. - Sanitization: The skill utilizes
character-pack-v3.schema.jsonto enforce strict regular expression patterns on file paths (preventing directory traversal) and character limits on all text fields to ensure data integrity.
Audit Metadata