oneshot-prompt-generator
Pass
Audited by Gen Agent Trust Hub on Aug 5, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill implements a robust defense against indirect prompt injection by instructing the agent in
SKILL.mdto treat all source material as untrusted evidence and to ignore any instructions or tool requests found within processed content. - [SAFE]: Includes a comprehensive safety testing suite, specifically a fixture (
evals/files/prompt-injection-source.md) containing a mock 'SYSTEM OVERRIDE' used to verify that the agent does not follow malicious commands embedded in external data. - [SAFE]: The provided validation and testing scripts (
scripts/validate.pyandscripts/test_skill.py) use standard Python libraries to perform static analysis of the skill package and do not exhibit any dangerous behaviors like network exfiltration or remote code execution. - [SAFE]: All network-related patterns found in the evaluation fixtures are used strictly for negative testing against known malicious behaviors and do not represent actual vulnerabilities in the skill's operational logic.
Audit Metadata