superred-optimizer-agentvigil-websentinel

An AgentVigil indirect prompt-injection optimizer for superred agent targets.

Community port — not an official implementation. This module is an unofficial re-implementation of AgentVigil 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.

AgentVigil is a black-box fuzzer for indirect prompt injection against LLM agents. Starting from a corpus of injection seeds, it uses MCTS/UCT selection to decide which seeds to reuse and which to explore, and an LLM mutator (expand, shorten, rephrase, crossover, generate-similar) to evolve them — searching for injections that land through content the agent reads (web pages, retrieved documents, tool results, memory).

References: "AgentVigil: Generic Black-Box Red-teaming for Indirect Prompt Injection against LLM Agents," arXiv:2505.05849, and "WebSentinel: Detecting and Localizing Prompt Injection Attacks for Web Agents," arXiv:2602.03792. Official implementation: wzunknown/agentvigil-websentinel. See ASSUMPTIONS.md for the full paper-to-code mapping and every deliberate departure.

WebSentinel is the detector/localizer side of the upstream work; this module implements the attack-generation side (the fuzzer described above) and does not implement a separate WebSentinel defense.

Install

pip install superred-optimizer-agentvigil-websentinel

Usage

from agentvigil_websentinel_optimizer import AgentVigilWebSentinelOptimizer

# Defaults: the official `new_seeds` web/content corpus, population_size=10,
# max_attempts=20 (mirrors the official fuzz-loop count).
optimizer = AgentVigilWebSentinelOptimizer()

# Also include the older official text-seed corpus.
optimizer = AgentVigilWebSentinelOptimizer(include_text_seeds=True)

This optimizer targets agent-style targets (AgentDojo, browser/RAG/memory/ MCP agents) that expose a content or tool-result surface. Wire it into a Controller like any other superred optimizer — see the framework's architecture overview.

License

This module is licensed under the MIT License.

It bundles seed corpora and mutator prompt text extracted from the upstream reference implementation, wzunknown/agentvigil-websentinel (the mutator prompts from mutate_prompts.py and the seed corpora from new_seeds.py/seeds.py). That repository publishes no license file; the upstream authors granted the superred maintainers permission, in private correspondence, to redistribute the ported material under MIT terms. See NOTICE for the full attribution and LICENSES/ for the license text applied to the upstream material.

Key dates

PyPI data
Data sourced directly from PyPI's database.
  • Released:
Latest release

1 maintainer

PyPI data
Data sourced directly from PyPI's database.
Avatar for simonissure from gravatar.com simonissure

Credits

Author: Sarthak Munshi

License expression

MIT
View SPDX License List

Requires

Python <3.14, >=3.11

Provides Extra

test

Report project as malware

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

superred_optimizer_agentvigil_websentinel-0.1.0.tar.gz (43.1 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

superred_optimizer_agentvigil_websentinel-0.1.0-py3-none-any.whl (30.7 kB view details)

Uploaded Python 3

File details

Details for the file superred_optimizer_agentvigil_websentinel-0.1.0.tar.gz.

File metadata

File hashes

Hashes for superred_optimizer_agentvigil_websentinel-0.1.0.tar.gz
Algorithm Hash digest
SHA256 73631d54bec1b3bb26788cd0c505163fef496967a66bc2ee55c11c81184ce0e1
MD5 64b1adb304615154211f75528ddc967e
BLAKE2b-256 6cdf1509fa7a7fb7467d491ae774dc294bf01740b8d11c9e38ec07931d075646

See more details on using hashes here.

Provenance

The following attestation bundles were made for superred_optimizer_agentvigil_websentinel-0.1.0.tar.gz:

Publisher: release.yml on RoldSI/superred-modules

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file superred_optimizer_agentvigil_websentinel-0.1.0-py3-none-any.whl.

File metadata

File hashes

Hashes for superred_optimizer_agentvigil_websentinel-0.1.0-py3-none-any.whl
Algorithm Hash digest
SHA256 80c3924b3631d7192fde487b7bf605f97bf466f46764d94a0ac7c2187c05c6c2
MD5 9ba4ea877a231835f6c35931b167372c
BLAKE2b-256 4ae31ce0db2071954a1f69041d986c9db3d58ff42444da37020e6cc77486bbab

See more details on using hashes here.

Provenance

The following attestation bundles were made for superred_optimizer_agentvigil_websentinel-0.1.0-py3-none-any.whl:

Publisher: release.yml on RoldSI/superred-modules

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

Release history Release notifications | RSS feed

This release

0.1.0 This release

2 files

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page