lov-laozhang-style-media-creator
Pass
Audited by Gen Agent Trust Hub on Oct 24, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is a comprehensive toolset for style-consistent media generation. It demonstrates high security standards through the inclusion of extensive measurement data, schema-based validation, and clear integrity guidelines in the
references/directory. - [COMMAND_EXECUTION]: The
scripts/analyze_media.pyutility invokes local media tools (ffmpegandffprobe) to extract technical metrics from video files. These operations are restricted to local paths and are standard for the skill's media analysis functionality, posing no risk of arbitrary command injection or privilege escalation. - [DATA_EXPOSURE_&_EXFILTRATION]: User profile management in
scripts/profile_store.pyincludes a proactive security layer that prevents the persistence of sensitive identifiers such as tokens, passwords, or API keys. The skill is designed for local-first use and does not perform unauthorized network operations. - [INDIRECT_PROMPT_INJECTION]: The skill ingests user-provided briefs and reference data. This potential attack surface is effectively mitigated by strict schema validation (
brief.schema.json,edit-plan.schema.json) and specific instructions in thereferences/authorship-integrity.mdcontract, which requires the agent to distinguish between facts, fictional demonstrations, and interpretations.
Audit Metadata