What functions / features it might have (expected capabilities)

Based on what one would expect from such a tool, here are features it might support (or need to support) if it is well‑designed:

1. Decompression (reading)

   • Accept a .zst or .savestate.zst file, decompress the Zstandard layer.

   • Parse any additional headers or metadata that wrap around the compressed payload (e.g. version tags, checksums, signature, identifier, offsets).

   • Output the raw savestate data (memory snapshot, registers, etc.) in a usable format.

2. Compression (writing)

   • Take a raw savestate data blob (perhaps after editing or patching), and compress it using Zstandard to .zst format.

   • Reconstruct or re-insert necessary headers/metadata so that the resulting .zst file is accepted by whatever system uses it (emulator, console, etc.).

   • Optionally allow configuration of compression levels, dictionary usage, multi-threading, etc.

3. Validation / checks

   • Validate the integrity of the file (e.g. confirm checksums, verify original size or magic values).

   • Possibly allow comparing the recompressed output against the original to check for equivalence (though compression is seldom exactly reversible bit-for-bit, unless it’s done with identical settings)