think-affinity-mapping
Pass
Audited by Gen Agent Trust Hub on Nov 9, 2026
Risk Level: SAFENO_CODEINDIRECT_PROMPT_INJECTION
Full Analysis
- [NO_CODE]: The skill is composed entirely of Markdown documentation and templates. There are no executable scripts, binaries, or shell commands included in the package.
- [SAFE]: Analysis across all threat categories confirmed the absence of malicious patterns such as credential harvesting, command execution, or obfuscated payloads.
- [INDIRECT_PROMPT_INJECTION]: The skill defines a process for analyzing external items which may contain untrusted instructions. 1. Ingestion points: User-provided text items (referenced in SKILL.md and TEMPLATE.md). 2. Boundary markers: No specific delimiters or safety warnings are provided for the input data. 3. Capability inventory: No subprocess calls, network operations, or file system modifications are used or available. 4. Sanitization: The skill does not perform validation or filtering of the processed items. Assessment: The risk is negligible as the skill lacks any tools or execution environment to exploit.
Audit Metadata