Paper Citation Record · LEDGER

Invisible Ink Threats: Adversarial Goals Behind Legitimate Tasks in Computer-Use Agents

As of 9 August 2026, this Paper Citation Record lists 15 of 15 outbound references and 0 inbound Pith citation observations for arXiv:2608.02018.

A citation records a reference. It does not transfer a finding from one paper to another.

pith.paper-citation-record.v1
2608.02018 v2

Coverage vector

measured 15 of 15 reference resolution

Typed states for the displayed outbound observations.

Source: paper_references, paper_reference_links, observed 2026-08-07T00:59:40.856673Z

measured 15 of 15 standing notices

One-hop event checks from named stored sources.

Source: scholarly_work_events, retraction_status_cache, observed 2026-08-08T06:32:00.761636+00:00

measured 0 of 0 inbound itemization

Pith citing papers itemized under the disclosed page cap.

Source: paper_references, paper_reference_links

measured 0 of 1 external citation measurements

A source-named dated measurement, never combined with another source.

Source: cited_works

Reference resolution

15 of 15 outbound references displayed

verified exact2
verified fuzzy5
unresolved8
parse uncertain0
malformed identifier0
metadata mismatch0

External citation measurements

No source-named external measurement is stored.

Outbound references

Observation c6e545e5-2312-4aec-81e7-14b2b7e4ff67 · outbound

This paper cites The Obvious Invisible Threat: LLM-Powered GUI Agents' Vulnerability to Fine-Print Injections.

Invisible Ink Threats: Adversarial Goals Behind Legitimate Tasks in Computer-Use Agents The Obvious Invisible Threat: LLM-Powered GUI Agents' Vulnerability to Fine-Print Injections

Reference 2

Resolution: unresolved; no resolver link, observed 2026-08-07T00:59:39.211103Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation fc0d42f9-143a-407f-b676-7447d1e75459 · outbound

This paper cites Defeating Prompt Injections by Design.

Invisible Ink Threats: Adversarial Goals Behind Legitimate Tasks in Computer-Use Agents Defeating Prompt Injections by Design

Reference 3

Resolution: unresolved; no resolver link, observed 2026-08-07T00:59:39.344088Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 08299e70-a016-4e1f-bdb9-6758bdf1aeb1 · outbound

This paper cites MiniMax Sparse Attention.

Invisible Ink Threats: Adversarial Goals Behind Legitimate Tasks in Computer-Use Agents MiniMax Sparse Attention

Reference 6

Resolution: verified exact; local_arxiv, observed 2026-08-07T00:59:41.192533Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation f9c7f59c-9fc1-4e6e-b474-ad1fb264f142 · outbound

This paper cites Commercial LLM Agents Are Already Vulnerable to Simple Yet Dangerous Attacks.

Invisible Ink Threats: Adversarial Goals Behind Legitimate Tasks in Computer-Use Agents Commercial LLM Agents Are Already Vulnerable to Simple Yet Dangerous Attacks

Reference 7

Resolution: unresolved; no resolver link, observed 2026-08-07T00:59:39.874950Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 89b4951e-49b9-498b-b772-1d29271513d9 · outbound

This paper cites Preference Redirection via Attention Concentration: An Attack on Computer Use Agents.

Invisible Ink Threats: Adversarial Goals Behind Legitimate Tasks in Computer-Use Agents Preference Redirection via Attention Concentration: An Attack on Computer Use Agents

Reference 9

Resolution: unresolved; no resolver link, observed 2026-08-07T00:59:40.104980Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation eb1eaa77-93ae-4b68-aa3b-af0bbdc6e997 · outbound

This paper cites Progent: Securing AI Agents with Privilege Control.

Invisible Ink Threats: Adversarial Goals Behind Legitimate Tasks in Computer-Use Agents Progent: Securing AI Agents with Privilege Control

Reference 10

Resolution: unresolved; no resolver link, observed 2026-08-07T00:59:40.195526Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation bb09119b-add7-40e4-aa93-454e0da504d3 · outbound

This paper cites OpenAI GPT-5 System Card.

Invisible Ink Threats: Adversarial Goals Behind Legitimate Tasks in Computer-Use Agents OpenAI GPT-5 System Card

Reference 11

Resolution: unresolved; no resolver link, observed 2026-08-07T00:59:40.324404Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 68d6e4cf-ab8c-4183-b43e-f310b052f6e6 · outbound

This paper cites InFindings of the Association for Computational Linguistics: ACL 2024, 10471–10506.

Invisible Ink Threats: Adversarial Goals Behind Legitimate Tasks in Computer-Use Agents InFindings of the Association for Computational Linguistics: ACL 2024, 10471–10506

Reference 14

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T00:59:42.133925Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation a4234450-0f60-41cd-9218-02b95c1b94a7 · outbound

This paper cites InProceedings of the 63rd Annual Meeting of the Association for Compu- tational Linguistics (Volume 1: Long Papers), 8387–8401.

Invisible Ink Threats: Adversarial Goals Behind Legitimate Tasks in Computer-Use Agents InProceedings of the 63rd Annual Meeting of the Association for Compu- tational Linguistics (Volume 1: Long Papers), 8387–8401

Reference 15

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T00:59:41.858013Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation 19deeb7a-d794-4f53-924a-7747ea87244c · outbound

This paper cites InFindings of the Association for Computa- tionalLinguistics:ACL2026,11986–11998.SanDiego,Cal- ifornia, United States: Association for Computational Lin- guistics.

Invisible Ink Threats: Adversarial Goals Behind Legitimate Tasks in Computer-Use Agents InFindings of the Association for Computa- tionalLinguistics:ACL2026,11986–11998.SanDiego,Cal- ifornia, United States: Association for Computational Lin- guistics

Reference 38

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T00:59:42.461591Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation 2090079e-44f3-46cf-9a7a-a44fab98e8a6 · outbound

This paper cites InProceed- ings of the 30th ACM International Conference on Multime- dia,MM’22,3185–3194.NewYork,NY,USA:Association for Computing Machinery.

Invisible Ink Threats: Adversarial Goals Behind Legitimate Tasks in Computer-Use Agents InProceed- ings of the 30th ACM International Conference on Multime- dia,MM’22,3185–3194.NewYork,NY,USA:Association for Computing Machinery

Reference 2022

Resolution: unresolved; no resolver link, observed 2026-08-07T00:59:40.004103Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation a7341e7a-89d0-46d2-bb5d-f5c3ab5033d3 · outbound

This paper cites InFindings of the Association for Computational Linguistics: EMNLP 2023, 2803–2821.

Invisible Ink Threats: Adversarial Goals Behind Legitimate Tasks in Computer-Use Agents InFindings of the Association for Computational Linguistics: EMNLP 2023, 2803–2821

Reference 2023

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T00:59:43.212515Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation a4c2b143-0641-481b-906c-feaccfe668e1 · outbound

This paper cites InAdvances in Neural Information Processing Systems, volume 37, 52040–52094.

Invisible Ink Threats: Adversarial Goals Behind Legitimate Tasks in Computer-Use Agents InAdvances in Neural Information Processing Systems, volume 37, 52040–52094

Reference 2024

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T00:59:42.849898Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation 8badb16f-e5e7-4198-8884-3c24f3e7d17b · outbound

This paper cites arXiv:2504.14064.

Invisible Ink Threats: Adversarial Goals Behind Legitimate Tasks in Computer-Use Agents arXiv:2504.14064

Reference 2025

Resolution: unresolved; no resolver link, observed 2026-08-07T00:59:39.120990Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 5c8e3839-ae4d-44ef-9dfe-7254b92435b5 · outbound

This paper cites MIRAGE: Context-Aware Prompt Injection against Mobile GUI Agents via User-Generated Content.

Invisible Ink Threats: Adversarial Goals Behind Legitimate Tasks in Computer-Use Agents MIRAGE: Context-Aware Prompt Injection against Mobile GUI Agents via User-Generated Content

Reference 2026

Resolution: verified exact; local_arxiv, observed 2026-08-07T00:59:41.483092Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Pith citing papers

No inbound Pith citation observations are available.