needscope

Pass

Audited by Gen Agent Trust Hub on Jul 14, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill follows security best practices by implementing a structured workflow that includes 'Gate' rules for sanitizing and validating data. It explicitly separates untrusted external data (social media comments) from internal processing logic using 'Contract Chains' and quality audits.
  • [COMMAND_EXECUTION]: The repository contains a Node.js validation script (scripts/validate_contract_chain.mjs) and a legacy Python report generator (references/archive/generate_report_html.legacy.py). Analysis confirms these scripts perform standard local file system operations and data processing without unsafe command injection, remote code execution, or network exfiltration.
  • [PROMPT_INJECTION]: The skill is designed to process untrusted social media text. However, it incorporates specific 'Object-Related Filters' and 'Quality Gates' (Step 3 and Step 5) to categorize and filter input data, effectively mitigating the risk of indirect prompt injection by ensuring that only verified, brand-related signals influence the final AI analysis and report generation.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 14, 2026, 07:38 AM
Security Audit — agent-trust-hub — needscope