style-detector
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFENO_CODE
Full Analysis
- No Code (SAFE): The skill consists entirely of Markdown files providing instructions and reference data for the agent. It does not include Python, JavaScript, or any other executable scripts, eliminating the risk of malware or remote code execution.
- Data Exposure & Exfiltration (SAFE): The skill utilizes the 'Read' and 'Edit' tools to manage style templates and project specifications locally. No hardcoded credentials, sensitive system paths, or network exfiltration patterns (e.g., curl, fetch) were detected.
- Prompt Injection (SAFE): Analysis of the instruction set shows no attempts to override agent safety guidelines, bypass constraints, or extract system prompts. The use of keywords for style detection is a standard functional requirement for the intended purpose.
- Indirect Prompt Injection (SAFE): While the skill ingests user input to determine writing styles, the attack surface is minimal as the input is used to choose from a predefined list of internal templates.
- Ingestion points: User input containing style keywords processed via KEYWORDS.md.
- Boundary markers: None explicitly defined.
- Capability inventory: Read (styles folder) and Edit (specification.md).
- Sanitization: The agent acts as the intermediary, selecting specific predefined paths based on keyword matching.
- Obfuscation (SAFE): No encoded strings (Base64), zero-width characters, or homoglyph-based evasion techniques were found in any of the skill files.
Audit Metadata