novel-release-packaging
Pass
Audited by Gen Agent Trust Hub on Mar 3, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [NO_CODE]: The skill consists entirely of Markdown documentation and templates. No executable Python, JavaScript, or shell scripts are included, eliminating the risk of direct command execution or malware.
- [PROMPT_INJECTION]: The instructions focus on formatting, noun consistency, and sensitivity checks. No patterns resembling system prompt overrides, safety filter bypasses, or 'jailbreak' attempts were identified.
- [DATA_EXFILTRATION]: No network-enabled tools or commands (such as curl, wget, or fetch) are present. The skill does not access sensitive system paths or hardcoded credentials.
- [REMOTE_CODE_EXECUTION]: There are no external package dependencies or mechanisms to download and execute remote scripts.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process draft chapters provided by the user. While this creates an ingestion point for untrusted data, the skill lacks the necessary capabilities (network access, file system modification, or command execution) to act on malicious instructions embedded in those drafts.
Audit Metadata