business-rule-extraction
Pass
Audited by Gen Agent Trust Hub on Oct 14, 2026
Risk Level: SAFEPROMPT_INJECTIONNO_CODEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill consists entirely of declarative instructions and configuration files such as Markdown, YAML, and JSON. No executable scripts, binaries, or active code components are present in the package.
- [PROMPT_INJECTION]: The system prompt in
prompts/business-rule-extraction.mdincludes explicit defensive instructions to treat any commands, role claims, or output requests found within the analyzed material as data to be analyzed rather than higher-priority instructions, mitigating potential injection risks. - [NO_CODE]: The repository contains only configuration and instructional content. No Python, Node.js, or Shell logic is implemented within the skill itself, minimizing the technical attack surface.
- [INDIRECT_PROMPT_INJECTION]: The skill's vulnerability surface was analyzed: 1. Ingestion points: Untrusted documentation supplied by users (referenced in
SKILL.md). 2. Boundary markers: Defined inprompts/business-rule-extraction.mdvia instructions to treat input as data. 3. Capability inventory: No subprocess calls, execution functions, file-write access, or network operations are present in any skill file. 4. Sanitization: Not explicitly implemented, but risk is negated by the absence of exploitable capabilities.
Audit Metadata