writing-dna-discovery
Pass
Audited by Gen Agent Trust Hub on Jun 13, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns detected. The skill's instructions and reference materials are focused entirely on the legitimate analysis of writing styles.
- [SAFE]: No obfuscated content, such as Base64-encoded commands or hidden Unicode characters, was found in any of the provided files.
- [SAFE]: The skill does not use any external dependencies, download remote code, or perform any network operations. All operations are local to the agent's conversational context.
- [SAFE]: Indirect Prompt Injection Surface. While the skill ingests untrusted user writing samples as described in 'SKILL.md', it lacks the capabilities required for this to be a security risk. Evidence chain:
- Ingestion points: User-provided writing samples are processed during the 'Initial Scan' phase in SKILL.md.
- Boundary markers: No explicit delimiters are defined to separate user samples from the agent's instructions.
- Capability inventory: No system commands, network access, or external tool invocations are present in the skill's logic.
- Sanitization: No sanitization is performed on input samples. However, given the lack of dangerous tools, the impact of a potential injection is restricted to the generated 'Voice DNA' document and does not affect the host system's security.
Audit Metadata