creativeclaw-edit-media
Pass
Audited by Gen Agent Trust Hub on Oct 13, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill and its referenced documentation do not contain any detected malicious patterns. No evidence of prompt injection, data exfiltration, or obfuscation was found. All infrastructure references point to the vendor's own domain or common public placeholders.
- [INDIRECT_PROMPT_INJECTION]: The skill processes external media files and transcripts, representing a standard surface for indirect prompt injection.
- Ingestion points: External media files and URLs provided via
upload_asset,import_chatgpt_media, or directly in tool parameters as described inSKILL.mdandreferences/platform-upload.md. - Boundary markers: The instructions do not define specific delimiters or "ignore embedded instructions" warnings for the agent when processing untrusted media content.
- Capability inventory: The skill utilizes an MCP server providing tools for media transformation, merging, and transcription (e.g.,
trim_video,cut_and_reframe_video,merge_media). - Sanitization: The provided instructions emphasize technical verification of duration and framing but do not specify sanitization of potential instructions embedded in media metadata or transcripts.
Audit Metadata