Repair Range Request Corruption

The problem

Resumable downloads fail when byte offsets are off-by-one or range headers disagree with the actual payload. These bugs are notoriously difficult to debug, often resulting in corrupted files, duplicated data, or incorrect 206 Partial Content responses.

What it does

• Analyzes Range and Content-Range parsing logic to identify inclusive interval math errors.
• Fixes suffix, open-ended, and multipart range calculations that lead to boundary byte corruption.
• Ensures If-Range validators correctly prevent serving stale bytes from modified representations.
• Standardizes response framing to align headers with actual payload lengths.
• Generates deterministic local fixtures to reproduce and verify fixes without production access.

Why this beats prompting it yourself

General LLMs often struggle with the precise 0-indexed vs 1-indexed math required for HTTP range requests and frequently suggest "fixes" that break multipart compliance. This skill enforces a strict ownership and normalization register, ensuring that authorization and cache policies remain untouched while the underlying byte math is repaired.

Use cases

• Fixing corrupted file downloads where the first or last bytes are missing.
• Correcting 416 Requested Range Not Satisfiable errors on valid files.
• Repairing multipart range responses where boundaries are incorrectly merged.
• Synchronizing compressed offsets with uncompressed file bounds.

Known limitations

Does not modify authorization logic, storage drivers, or external retry policies. Requires local repository files to function and cannot test against live production endpoints.