commercial-proposal-writing
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFE
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill includes several Python scripts for statistical analysis (ANOVA, KMeans, Logistic Regression, etc.). These scripts are static, reside within the skill's own directory, and utilize standard, reputable libraries like Pandas, Scikit-learn, and Statsmodels. No evidence of fetching or executing remote payloads was found.
- [DATA_EXFILTRATION]: No network operations or commands to send data to external domains were identified. The skill focuses on processing user-provided business data and generating text or local analysis results.
- [PROMPT_INJECTION]: The skill uses strong instructional language and specific triggers to guide the AI's behavior for proposal writing quality. No patterns attempting to bypass safety filters or override system-level instructions were found.
- [DYNAMIC_EXECUTION]: While the skill utilizes Python scripts for analysis, these are predefined templates and do not involve dynamic code generation, eval(), or unsafe deserialization of untrusted data.
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied proposal drafts and evidence. It implements a 'Data Sufficiency Gate' in SKILL.md and multiple 'Quality Gates' (e.g., in 08-output-polish-and-pitch-checklist.md) to validate inputs and outputs, effectively managing the risk of unintended behavior from processed content.
Audit Metadata