clean-company-names-for-copy
Pass
Audited by Gen Agent Trust Hub on Oct 1, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill is designed for data cleaning and does not perform any high-risk operations such as arbitrary command execution, privilege escalation, or persistence mechanisms.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input strings from company names and domains.
- [INDIRECT_PROMPT_INJECTION]: Ingestion points: The skill ingests data from company_name_raw and domain fields provided in a table or CSV (SKILL.md, Step 0 and Step 3).
- [INDIRECT_PROMPT_INJECTION]: Boundary markers: The skill mitigates injection risks by using a locked prompt and requiring a JSON response format for the model output (SKILL.md, Step 3).
- [INDIRECT_PROMPT_INJECTION]: Capability inventory: The skill's capabilities are restricted to reading and writing specific columns within the platform's table environment; it does not possess network exfiltration or shell access capabilities.
- [INDIRECT_PROMPT_INJECTION]: Sanitization: A mandatory placeholder guard (blocklist) is applied to both inputs and outputs to filter out junk strings and common placeholders, preventing them from being processed as brands (SKILL.md, Step 3).
Audit Metadata