Paper Citation Record · LEDGER

AgentSecBench: Measuring Prompt Injection, Privacy Leakage, and Tool-Use Integrity in LLM Agents

As of 21 August 2026, this Paper Citation Record lists 18 of 18 outbound references and 2 inbound Pith citation observations for arXiv:2605.26269.

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

pith.paper-citation-record.v1
2605.26269 v1

Coverage vector

measured 18 of 18 reference resolution

Typed states for the displayed outbound observations.

Source: paper_references, paper_reference_links, observed 2026-06-29T21:16:48.677159Z

measured 20 of 20 standing notices

One-hop event checks from named stored sources.

Source: scholarly_work_events, retraction_status_cache, observed 2026-08-20T06:33:59.587034+00:00

measured 2 of 2 inbound itemization

Pith citing papers itemized under the disclosed page cap.

Source: paper_references, paper_reference_links, observed 2026-08-15T22:04:21.540541Z

measured 0 of 1 external citation measurements

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

Source: pith, observed 2026-07-04T14:09:53.283015Z

Reference resolution

18 of 18 outbound references displayed

verified exact0
verified fuzzy0
unresolved18
parse uncertain0
malformed identifier0
metadata mismatch0

External citation measurements

No source-named external measurement is stored.

Outbound references

Observation 72b62889-3a51-4557-a736-87959e31ac3c · outbound

This paper cites Tensor Trust: Interpretable prompt injection attacks from an online game,.

AgentSecBench: Measuring Prompt Injection, Privacy Leakage, and Tool-Use Integrity in LLM Agents Tensor Trust: Interpretable prompt injection attacks from an online game,

Reference 1

Resolution: unresolved; no resolver link, observed 2026-06-29T21:16:48.677159Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f1d684b3-ff6f-45da-9fa1-3efc333aa256 · outbound

This paper cites AgentDojo: A dynamic environment to evaluate prompt injection at- tacks and defenses for LLM agents,.

AgentSecBench: Measuring Prompt Injection, Privacy Leakage, and Tool-Use Integrity in LLM Agents AgentDojo: A dynamic environment to evaluate prompt injection at- tacks and defenses for LLM agents,

Reference 2

Resolution: unresolved; no resolver link, observed 2026-06-29T21:16:48.677159Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 3975ac22-6c29-4888-a344-8290da4fc525 · outbound

This paper cites Retrieval-augmented generation for knowledge-intensive NLP tasks,.

AgentSecBench: Measuring Prompt Injection, Privacy Leakage, and Tool-Use Integrity in LLM Agents Retrieval-augmented generation for knowledge-intensive NLP tasks,

Reference 3

Resolution: unresolved; no resolver link, observed 2026-06-29T21:16:48.677159Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 3fa95d3e-3aa2-47a8-a837-7e5eb2826c52 · outbound

This paper cites Dense passage retrieval for open-domain question answering,.

AgentSecBench: Measuring Prompt Injection, Privacy Leakage, and Tool-Use Integrity in LLM Agents Dense passage retrieval for open-domain question answering,

Reference 4

Resolution: unresolved; no resolver link, observed 2026-06-29T21:16:48.677159Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation ba7c5521-56d3-4223-b1e1-fdef7ca3e7bd · outbound

This paper cites Extracting training data from large language models,.

AgentSecBench: Measuring Prompt Injection, Privacy Leakage, and Tool-Use Integrity in LLM Agents Extracting training data from large language models,

Reference 5

Resolution: unresolved; no resolver link, observed 2026-06-29T21:16:48.677159Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation a7f658cd-2c68-4b6d-a6e9-389750daa030 · outbound

This paper cites ReAct: Synergizing reasoning and acting in language models,.

AgentSecBench: Measuring Prompt Injection, Privacy Leakage, and Tool-Use Integrity in LLM Agents ReAct: Synergizing reasoning and acting in language models,

Reference 6

Resolution: unresolved; no resolver link, observed 2026-06-29T21:16:48.677159Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation e57f61ee-7138-477f-b2c5-7621c50bba50 · outbound

This paper cites Toolformer: Language models can teach themselves to use tools,.

AgentSecBench: Measuring Prompt Injection, Privacy Leakage, and Tool-Use Integrity in LLM Agents Toolformer: Language models can teach themselves to use tools,

Reference 7

Resolution: unresolved; no resolver link, observed 2026-06-29T21:16:48.677159Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 7e3b1c49-756f-4c7f-bc1d-7f3763901688 · outbound

This paper cites Not what you’ve signed up for: Compromising real-world LLM-integrated applications with indirect prompt injection,.

AgentSecBench: Measuring Prompt Injection, Privacy Leakage, and Tool-Use Integrity in LLM Agents Not what you’ve signed up for: Compromising real-world LLM-integrated applications with indirect prompt injection,

Reference 8

Resolution: unresolved; no resolver link, observed 2026-06-29T21:16:48.677159Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 4730b1fe-5a19-43f9-b2a5-42b3ac01d75e · outbound

This paper cites Formalizing and benchmarking prompt injection attacks and defenses,.

AgentSecBench: Measuring Prompt Injection, Privacy Leakage, and Tool-Use Integrity in LLM Agents Formalizing and benchmarking prompt injection attacks and defenses,

Reference 9

Resolution: unresolved; no resolver link, observed 2026-06-29T21:16:48.677159Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 601e5bf5-1fa1-4fbc-ba71-c35c6d366634 · outbound

This paper cites InjecAgent: Benchmarking indirect prompt injections in tool-integrated large language model agents,.

AgentSecBench: Measuring Prompt Injection, Privacy Leakage, and Tool-Use Integrity in LLM Agents InjecAgent: Benchmarking indirect prompt injections in tool-integrated large language model agents,

Reference 10

Resolution: unresolved; no resolver link, observed 2026-06-29T21:16:48.677159Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 8253c0bd-53d1-45c6-9e61-58270ede9c5f · outbound

This paper cites Security policies and security models,.

AgentSecBench: Measuring Prompt Injection, Privacy Leakage, and Tool-Use Integrity in LLM Agents Security policies and security models,

Reference 11

Resolution: unresolved; no resolver link, observed 2026-06-29T21:16:48.677159Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 35a6cd26-9e68-4dab-aace-1f410191ba1c · outbound

This paper cites Language-based information-flow security,.

AgentSecBench: Measuring Prompt Injection, Privacy Leakage, and Tool-Use Integrity in LLM Agents Language-based information-flow security,

Reference 12

Resolution: unresolved; no resolver link, observed 2026-06-29T21:16:48.677159Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation e1b540df-fe38-471e-9500-b732d30d7cf7 · outbound

This paper cites Universally composable security: A new paradigm for cryptographic protocols,.

AgentSecBench: Measuring Prompt Injection, Privacy Leakage, and Tool-Use Integrity in LLM Agents Universally composable security: A new paradigm for cryptographic protocols,

Reference 13

Resolution: unresolved; no resolver link, observed 2026-06-29T21:16:48.677159Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 33bc4937-7f92-45ae-ab9c-0c8904a95629 · outbound

This paper cites Universal adversarial triggers for attacking and analyzing NLP,.

AgentSecBench: Measuring Prompt Injection, Privacy Leakage, and Tool-Use Integrity in LLM Agents Universal adversarial triggers for attacking and analyzing NLP,

Reference 14

Resolution: unresolved; no resolver link, observed 2026-06-29T21:16:48.677159Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation e9bb1629-2aca-4b79-b6a3-c86ce8e9c520 · outbound

This paper cites Universal and transferable ad- versarial attacks on aligned language models,.

AgentSecBench: Measuring Prompt Injection, Privacy Leakage, and Tool-Use Integrity in LLM Agents Universal and transferable ad- versarial attacks on aligned language models,

Reference 15

Resolution: unresolved; no resolver link, observed 2026-06-29T21:16:48.677159Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f17b21ec-8260-49d2-9b9f-ea7eff8a7ad7 · outbound

This paper cites Quantifying memorization across neural language models,.

AgentSecBench: Measuring Prompt Injection, Privacy Leakage, and Tool-Use Integrity in LLM Agents Quantifying memorization across neural language models,

Reference 16

Resolution: unresolved; no resolver link, observed 2026-06-29T21:16:48.677159Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 81fe9754-35d7-467b-bddc-b31e57a79621 · outbound

This paper cites Deduplicating training data mitigates privacy risks in language models,.

AgentSecBench: Measuring Prompt Injection, Privacy Leakage, and Tool-Use Integrity in LLM Agents Deduplicating training data mitigates privacy risks in language models,

Reference 17

Resolution: unresolved; no resolver link, observed 2026-06-29T21:16:48.677159Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 83303f67-6044-4557-bbc1-a789279c0763 · outbound

This paper cites Training language models to follow instructions with human feedback,.

AgentSecBench: Measuring Prompt Injection, Privacy Leakage, and Tool-Use Integrity in LLM Agents Training language models to follow instructions with human feedback,

Reference 18

Resolution: unresolved; no resolver link, observed 2026-06-29T21:16:48.677159Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Pith citing papers

Observation 7fa13487-93ae-4a6e-a7c4-08c76100a160 · inbound

Agents That Know Too Much: A Data-Centric Survey of Privacy in LLM Agents cites this paper.

Agents That Know Too Much: A Data-Centric Survey of Privacy in LLM Agents AgentSecBench: Measuring Prompt Injection, Privacy Leakage, and Tool-Use Integrity in LLM Agents

Reference 5

Resolution: verified exact; local_arxiv, observed 2026-07-04T14:09:53.284810Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 9be6a542-8daa-471e-8f38-2277f3d7ba5f · inbound

Practice Makes Unsafe: Skill Misevolution in Self-Improving LLM Agents cites this paper.

Practice Makes Unsafe: Skill Misevolution in Self-Improving LLM Agents AgentSecBench: Measuring Prompt Injection, Privacy Leakage, and Tool-Use Integrity in LLM Agents

Reference 38

Resolution: unresolved; no resolver link, observed 2026-08-15T22:04:21.540541Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.