skills/zengrong233/awesome-ai-research-writing-skill/awesome-ai-research-writing/Gen Agent Trust Hub
awesome-ai-research-writing
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFENO_CODEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill is composed exclusively of Markdown and YAML instruction files that guide the AI's behavior through prompt templates. No executable scripts (.py, .js, .sh), configuration files for package managers, or binaries were found.
- [NO_CODE]: There are no source code files or automation scripts included in the package, which significantly reduces the risk of traditional remote code execution, persistence, or privilege escalation.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to analyze external untrusted artifacts such as PDF papers, LaTeX manuscripts, and source code. * Ingestion points: External documents and manuscripts described in
SKILL.mdandreferences/paper-reading.md. * Boundary markers: The skill uses explicit 'Output Contracts' inSKILL.mdto scope the AI's behavior, though no formal sanitization is applied to input contents. * Capability inventory: Web browsing for fact-checking and structured reading of files. * Sanitization: Not applicable as the skill only provides text processing instructions. The risk is evaluated as safe as these capabilities are limited to the user's explicit research context. - [SAFE]: No obfuscation (Base64, zero-width characters, or homoglyphs) or metadata poisoning attempts were found. All URLs used in the documentation are relative local references.
Audit Metadata