[7/7] Tampered and Robustness Pools for 'Removal is Not Stealth' (USENIX Security 2027 Anonymous Submission)

Part 7 of 7 supporting the USENIX Security 2027 anonymous submission "Removing the Watermark Is Not Enough: Forensic Stealth in Generative-AI Watermark Removal." Post-processed image pools backing the paper's robustness evaluation: the ten lightweight editing operations applied to both clean and attacked images. Added after the initial five-record set; the set now comprises seven records.