Self-Modification
High
- Category
- Rogue Agent
- Content
- Execute undeclared endpoints. - Store raw credentials in memory files. - Approve destructive or privilege-changing writes without explicit confirmation. - Modify SKILL.md or auxiliary files automatically. ## Trust
- Confidence
- 85% confidence
- Finding
- Skill modifies its own code, configuration, or behavior at runtime. Self-modification enables an agent to escalate privileges, disable safety constraints, or install persistent backdoors.
