agmiAgent Memory Integrity GitHub

Front-door attack

Update poisoning

Write a 'correction' of a fact the user stated.

What it looks like in the world: Every real tool serves the correction alongside the original.

attack id update_poisoning · version 1 · attacker: write-access

Every store, this edit

surfaced 7

Reproduce

pip install agent-memory-integrity
python agmi/full_runner.py --json results/scorecard.json   # every row; the update_poisoning column is this page

All 15 edits and attacks · The scorecard