superred-optimizer-eia-agent
An EIA (Environmental Injection Attack) optimizer for superred web-agent targets.
Community port — not an official implementation. This module is an unofficial re-implementation of EIA (Liao et al., OSU NLP Group) for superred. It is not affiliated with, endorsed by, or maintained by the original authors. See ASSUMPTIONS.md for every deliberate deviation from the paper and reference code.
EIA attacks web agents by changing the environment rather than the instruction. It plants hidden or visually subtle HTML elements — persuasive labels, injected form fields, aria-labels — on a page the agent reads. The agent then grounds its next action to the attacker's element and types private information into it, leaking data without ever being told to.
Reference: Liao et al., "EIA: Environmental Injection Attack on Generalist Web
Agents for Privacy Leakage,"
arXiv:2409.11295 (ICLR 2025). Official
implementation:
OSU-NLP-Group/EIA_against_webagent.
See ASSUMPTIONS.md for the full paper-to-code mapping and every
deliberate departure.
Install
pip install superred-optimizer-eia-agent
Usage
from eia_agent_optimizer import EIAAgentOptimizer
# Defaults follow the paper/code mainline (action_grounding + form_type1,
# near_bot_1 placement, zero-opacity injection).
optimizer = EIAAgentOptimizer()
# Supply the leak target so the local no-feedback evaluator can verify the value.
optimizer = EIAAgentOptimizer(privacy_type="credit_card", target_secret="4111...")
This optimizer targets agent-style targets (AgentDojo-style browser/web
agents). Wire it into a Controller like any other superred optimizer — see the
framework's architecture overview.
License
MIT for this port's code (see LICENSE). The upstream reference
implementation,
OSU-NLP-Group/EIA_against_webagent,
is MIT licensed. The HTML form, style, and submit-script templates ported from
its injection/ code are credited in NOTICE, and its MIT copyright
notice is preserved in
LICENSES/EIA_against_webagent-MIT.txt.
(The separate SeeAct/ directory in that repository is under the AI PUBS
OpenRAIL-S license; this module does not use or copy from it — the bundled
templates come only from the MIT-licensed injection/ code.)
1 maintainer
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
File details
Details for the file superred_optimizer_eia_agent-0.1.0.tar.gz.
File metadata
- Download URL: superred_optimizer_eia_agent-0.1.0.tar.gz
- Upload date:
- Size: 34.0 kB
- Tags: Source
- Uploaded using Trusted Publishing? Yes
- Uploaded via:
twine/7.0.0 CPython/3.13.14
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
a206326e0c7899cb97dcf649a7244f1028069ab6a9295ddfdbfd85f14e8cb0ea
|
|
| MD5 |
9010aed634a3e72232c343a69302ac08
|
|
| BLAKE2b-256 |
2644a3a18ce96be9a01dd1a583d4d6b75f5bc9dd4f1f5e4853c5c53b7fb78cb5
|
Provenance
The following attestation bundles were made for superred_optimizer_eia_agent-0.1.0.tar.gz:
Publisher:
release.yml on RoldSI/superred-modules
-
Statement:
-
Statement type:
https://in-toto.io/Statement/v1 -
Predicate type:
https://docs.pypi.org/attestations/publish/v1 -
Subject name:
superred_optimizer_eia_agent-0.1.0.tar.gz -
Subject digest:
a206326e0c7899cb97dcf649a7244f1028069ab6a9295ddfdbfd85f14e8cb0ea - Sigstore transparency entry: 2519586260
- Sigstore integration time:
-
Permalink:
RoldSI/superred-modules@1c1cbc87f3969cd4417cb4ecbd685252ad82c185 -
Branch / Tag:
refs/tags/superred-optimizer-eia-agent-v0.1.0 - Owner: https://github.com/RoldSI
-
Access:
private
-
Token Issuer:
https://token.actions.githubusercontent.com -
Runner Environment:
github-hosted -
Publication workflow:
release.yml@1c1cbc87f3969cd4417cb4ecbd685252ad82c185 -
Trigger Event:
push
-
Statement type:
File details
Details for the file superred_optimizer_eia_agent-0.1.0-py3-none-any.whl.
File metadata
- Download URL: superred_optimizer_eia_agent-0.1.0-py3-none-any.whl
- Upload date:
- Size: 22.4 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? Yes
- Uploaded via:
twine/7.0.0 CPython/3.13.14
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
cea875370a868bcf784d4e12b5c2a86c00bf1b72f5e13cc0b1270c51d706b8c3
|
|
| MD5 |
ad05bf79bcbd2d24a1229cafe2108c73
|
|
| BLAKE2b-256 |
82b2cbd7eb22b88dbc41f03f500193e31df1fa3a858a9e801a252152590d1b4c
|
Provenance
The following attestation bundles were made for superred_optimizer_eia_agent-0.1.0-py3-none-any.whl:
Publisher:
release.yml on RoldSI/superred-modules
-
Statement:
-
Statement type:
https://in-toto.io/Statement/v1 -
Predicate type:
https://docs.pypi.org/attestations/publish/v1 -
Subject name:
superred_optimizer_eia_agent-0.1.0-py3-none-any.whl -
Subject digest:
cea875370a868bcf784d4e12b5c2a86c00bf1b72f5e13cc0b1270c51d706b8c3 - Sigstore transparency entry: 2519586828
- Sigstore integration time:
-
Permalink:
RoldSI/superred-modules@1c1cbc87f3969cd4417cb4ecbd685252ad82c185 -
Branch / Tag:
refs/tags/superred-optimizer-eia-agent-v0.1.0 - Owner: https://github.com/RoldSI
-
Access:
private
-
Token Issuer:
https://token.actions.githubusercontent.com -
Runner Environment:
github-hosted -
Publication workflow:
release.yml@1c1cbc87f3969cd4417cb4ecbd685252ad82c185 -
Trigger Event:
push
-
Statement type:
