seedance-vocab-en
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill's primary function is text transformation and vocabulary improvement. It provides a systematic approach to 'de-slopping' prompts by replacing empty quality adjectives (e.g., 'cinematic', 'epic') with observable production details (e.g., lens distance, lighting sources).
- [NO_CODE]: The skill consists entirely of Markdown instructions and YAML metadata. There are no executable scripts, binaries, or automated tool calls included in the provided file.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted user-supplied prompts (Ingestion points: SKILL.md description). It lacks explicit boundary markers for these inputs. However, the skill has no dangerous tools or automated actions (Capability inventory: No subprocess calls, file system writes, or network operations across the file). No sanitization is implemented as the skill's sole output is the transformed text (Sanitization: Absent).
- [PROMPT_INJECTION]: The skill includes a 'Filter-Aware Wording' section that provides synonyms for production terms that often trigger false positives in safety filters. This section explicitly states it is for clarity in safe prompts and is not intended for evasion of safety guidelines for genuinely risky content, which is routed to a dedicated filter skill.
Audit Metadata