skill-updater
Pass
Audited by Gen Agent Trust Hub on Feb 19, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [Prompt Injection] (SAFE): No instructions to override agent behavior or bypass safety guidelines were found.\n- [Data Exposure & Exfiltration] (SAFE): No credentials or sensitive file access patterns were detected.\n- [Obfuscation] (SAFE): No encoded or hidden content was found.\n- [Unverifiable Dependencies & Remote Code Execution] (SAFE): No remote scripts or external packages are downloaded.\n- [Privilege Escalation] (SAFE): No privilege escalation commands (like sudo) are used.\n- [Persistence Mechanisms] (SAFE): No persistence mechanisms were identified.\n- [Indirect Prompt Injection] (LOW): The skill's workflow involves ingesting untrusted data which could lead to instruction injection.\n
- Ingestion points: Improvement recommendations and user feedback gathered in Step 1 (SKILL.md).\n
- Boundary markers: Not used in the instructions.\n
- Capability inventory: Read, Write, Edit, Glob, Grep, Bash, WebSearch, WebFetch.\n
- Sanitization: Not specified in the implementation.
Audit Metadata