list-quality-scorecard
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute a local script to generate the scorecard. \n
- Evidence:
npx tsx scripts/score-list.ts --list=leads.csv --icp-file=client-profile.yaml --out=scorecard.mdinSKILL.md. \n - The script
scripts/score-list.tsuses standard Node.jsfsandprocessmodules to handle file I/O based on user arguments. \n- [INDIRECT_PROMPT_INJECTION]: The skill is designed to analyze external lead lists (CSV files) which could potentially contain malicious data meant to influence the agent. \n - Ingestion points:
leads.csvandclient-profile.yamlare read viareadFileSyncinscripts/score-list.ts. \n - Boundary markers: None present in CSV parsing logic. \n
- Capability inventory: The script has file read/write permissions (
readFileSync,writeFileSync) but no network or shell execution capabilities. \n - Sanitization: The script extracts specific data points (emails, titles, names) and uses them to compute numeric scores and match against predefined patterns. It does not echo raw row content into the final markdown report except for file metadata. \n- [SAFE]: The analysis found no evidence of malicious behavior such as credential harvesting, unauthorized network communication, obfuscation, or persistence mechanisms. The skill's operations are confined to local data processing.
Audit Metadata