ton-auditor

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.90). The skill explicitly fetches external content: SKILL.md Turn 1 runs a curl against raw.githubusercontent.com for a remote VERSION, and assets/docs/README.md states "To reference online docs, create a file containing one URL per line — they will be fetched and read automatically," meaning arbitrary public URLs can be fetched and read as part of the audit workflow and thus could inject instructions that influence agent decisions.