FOITW pre-registration catalog
The public view of the catalog described in the FOITW methodology. Each entry names the Threat Matrix technique a signature watches for, the published research it derives from, when we registered it, and whether it has fired.
What this view shows, and what it does not
- Each registration time is our own record. It is a commit in our research repository, which is not public, so a reader cannot check it independently.
- The signatures are not published: not the match expression, not the field it reads, and not a signature identifier. Nothing on this page binds the signature that fires to the signature that was registered.
- Anchoring registrations in the registry transparency log is not yet implemented, so no entry carries an anchor.
- What a reader can check is that each entry, with its technique and its source, is listed publicly. That listing dates from the first public copy of this page, such as an independent archive capture, and not from the registration time.
Generated from the catalog at research repository commit 433223875d84c37329a73d8d47db928d574493f4.
Entries
FOITW-CAT-0001
- Technique
- T-2002
- Source
- Not what you've signed up for: Compromising Real-World LLM-Integrated Applications with Indirect Prompt Injection (AISec '23)Greshake et al.Published
- Registered
- Transparency log
- Not anchored
- Status
No fire recorded
The catalog at commit
433223875d84records no fire for this entry. That is the absence of a record, not a measurement that the technique has not been used.
FOITW-CAT-0002
- Technique
- T-2007
- Source
- Lessons from Defending Gemini Against Indirect Prompt Injections (Google DeepMind, 2025)Shi, Lin, Song, Hayes, Shumailov, Yona, Pluto, Pappu, Choquette-Choo, Nasr, Sitawarin, Gibson, Terzis, FlynnPublished
- Registered
- Transparency log
- Not anchored
- Status
No fire recorded
The catalog at commit
433223875d84records no fire for this entry. That is the absence of a record, not a measurement that the technique has not been used.
FOITW-CAT-0003
- Technique
- T-9005
- Source
- AI threats in the wild: The current state of prompt injections on the web (Google Online Security Blog, 2026)Brunner, Liu, PandePublished
- Registered
- Transparency log
- Not anchored
- Status
No fire recorded
The catalog at commit
433223875d84records no fire for this entry. That is the absence of a record, not a measurement that the technique has not been used.