Pith. sign in

Paper Citation Record · LEDGER

Agent Safety Should Be a Runtime Contract

As of 19 August 2026, this Paper Citation Record lists 100 of 103 outbound references and 0 inbound Pith citation observations for arXiv:2608.11274.

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

pith.paper-citation-record.v1
2608.11274 v1

Coverage vector

measured 100 of 103 reference resolution

Typed states for the displayed outbound observations.

Source: paper_references, paper_reference_links, observed 2026-08-15T14:19:14.004821Z

measured 100 of 100 standing notices

One-hop event checks from named stored sources.

Source: scholarly_work_events, retraction_status_cache, observed 2026-08-18T06:34:40.430872+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

100 of 103 outbound references displayed

  • verified exact3
  • verified fuzzy32
  • unresolved64
  • parse uncertain0
  • malformed identifier0
  • metadata mismatch1

External citation measurements

No source-named external measurement is stored.

Outbound references

Observation 449d71cd-af24-4bce-905e-3d0f767b94e5 · outbound

This paper cites Incident 622: Chevrolet dealer chatbot agrees to sell tahoe for $1, 2023.

Agent Safety Should Be a Runtime Contract Incident 622: Chevrolet dealer chatbot agrees to sell tahoe for $1, 2023

Reference 1

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.363913Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.363913Z digest=sha256:4792f7a196ee402c0812d23e79a4ed5aa4cf5f42622e9063cf1803d33632f735

Observation c1ab0b82-12c0-4dd7-8dbe-f8d9480753df · outbound

This paper cites Incident 631: DPD chatbot malfunctioned and swore at customers, 2024.

Agent Safety Should Be a Runtime Contract Incident 631: DPD chatbot malfunctioned and swore at customers, 2024

Reference 2

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.370084Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.370084Z digest=sha256:5eac4c568867297774bd1b5548b29c50fca4a73d29af57f07bf2e4215842abc0

Observation f83fd8b2-e21e-482c-ba39-757c2b287e9d · outbound

This paper cites Incident 1152: LLM-driven Replit agent reportedly executed unauthorized destructive commands during code freeze, 2025.https://incidentdatabase.ai/cite/1152/.

Agent Safety Should Be a Runtime Contract Incident 1152: LLM-driven Replit agent reportedly executed unauthorized destructive commands during code freeze, 2025.https://incidentdatabase.ai/cite/1152/

Reference 3

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.379621Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.379621Z digest=sha256:188e1dfc0fb622e428cc1463fe4be6b0a628d27d9c8f2b40cca6fc3e4ee94cfd

Observation 843c32ab-8190-4b4b-b3cc-3251abeec668 · outbound

This paper cites Alur.Principles of Cyber-Physical Systems.

Agent Safety Should Be a Runtime Contract Alur.Principles of Cyber-Physical Systems

Reference 4

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.385430Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.385430Z digest=sha256:80be8ee1f1ed104007a25b4c968fb54bb67e37a362f42b992043b691ff99634d

Observation 481d0a6d-556c-438c-905b-f6755d7a2e07 · outbound

This paper cites Anderson.Security Engineering: A Guide to Building Dependable Distributed Systems.

Agent Safety Should Be a Runtime Contract Anderson.Security Engineering: A Guide to Building Dependable Distributed Systems

Reference 5

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.391141Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.391141Z digest=sha256:55537871f529b0c5034b0643568c8e24b35bec8be902d62f88c7e5d78779184e

Observation 11af2645-9bc0-481f-bd95-a89290be909f · outbound

This paper cites an unresolved cited work.

Agent Safety Should Be a Runtime Contract Unresolved cited work

Reference 6

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.396657Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.396657Z digest=sha256:d115585c29d2e7ec2656546e627beaa3840a14b93d7e5252323ae3f7c9ba0c94

Observation 8bd41e07-c458-4357-b017-1516a2d3e4e1 · outbound

This paper cites Claude 3.5 sonnet model card addendum.Anthropic Technical Report, 2024.

Agent Safety Should Be a Runtime Contract Claude 3.5 sonnet model card addendum.Anthropic Technical Report, 2024

Reference 7

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.404241Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.404241Z digest=sha256:00cadc668a11fe6a86758a0ab42a97a78d5eff2e534028caedacd98a1a4c3ece

Observation 787f9c8f-d96f-49e1-9d58-8bcceaa02996 · outbound

This paper cites Cursor CLI output format reference, 2025.

Agent Safety Should Be a Runtime Contract Cursor CLI output format reference, 2025

Reference 8

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.409327Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.409327Z digest=sha256:636e6c62e2ae8982cc5a3365dc8a32ea3189cbd4d210393cade9757cbaea8f43

Observation d896fe0c-1514-4bb3-b4f6-fa32e6d607bb · outbound

This paper cites Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback.

Agent Safety Should Be a Runtime Contract Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback

Reference 9

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.414870Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.414870Z digest=sha256:7cba8a7db96d9b9aadf36cfc5dcf6300815c070c72ba7d492dabe67b4ef6ddf2

Observation 32a8c30e-6f3f-40f5-ac86-a9d52f3b5156 · outbound

This paper cites Constitutional AI: Harmlessness from AI Feedback.

Agent Safety Should Be a Runtime Contract Constitutional AI: Harmlessness from AI Feedback

Reference 10

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.420823Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.420823Z digest=sha256:28d8c81ed70cc594b063bdfb16492d55cc93fc82a5f627d7882c366c1601d0e2

Observation d57e0c2d-52e1-46b6-93f8-2017cce3a708 · outbound

This paper cites Monitoring Reasoning Models for Misbehavior and the Risks of Promoting Obfuscation.

Agent Safety Should Be a Runtime Contract Monitoring Reasoning Models for Misbehavior and the Risks of Promoting Obfuscation

Reference 11

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.429309Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.429309Z digest=sha256:38f338853d1f7fa1064b9020eccf3b9fed22e8326bf197094856ec711f4155c0

Observation ea0cc557-e7d8-4bfe-959a-7d19bd1fc568 · outbound

This paper cites an unresolved cited work.

Agent Safety Should Be a Runtime Contract Unresolved cited work

Reference 12

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.435975Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.435975Z digest=sha256:67578c733f76fb9cc6b2e38efb93b7c421caafcfe3708d7103191fd4d70c4658

Observation 41e12a85-1b5d-44e1-b6cd-74288fe305f9 · outbound

This paper cites Bishop.Computer Security: Art and Science.

Agent Safety Should Be a Runtime Contract Bishop.Computer Security: Art and Science

Reference 13

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.442474Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.442474Z digest=sha256:cd5a554cc1275e9a6690cb896e3cea71917c8ce315b42d23b01814f9ba32d1d9

Observation c8bd34c8-0e5a-48c8-b8ae-576318e02a79 · outbound

This paper cites Bricken, A.

Agent Safety Should Be a Runtime Contract Bricken, A

Reference 14

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.448879Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.448879Z digest=sha256:86a589c5fd28e7d9825356348113423dd8d1a7496b7bc61a32b1754a88676fd4

Observation 93745e61-d63a-4e0a-b6b2-e0274edb82c9 · outbound

This paper cites Burke, H.

Agent Safety Should Be a Runtime Contract Burke, H

Reference 15

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.454815Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.454815Z digest=sha256:9f40c7f045b3c764b1e2a5e45c53464aee4b9735a4c620749adbd8a05183e32d

Observation 28c0d8d6-fb6c-4710-8276-c2b504988cc2 · outbound

This paper cites Carlini, M.

Agent Safety Should Be a Runtime Contract Carlini, M

Reference 16

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.460522Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.460522Z digest=sha256:d05c913e88d137c76b0641e09dd46a77bad5c7abd41afc0f39dbab87d8466402

Observation 19ff294a-5e4e-4f87-8bd5-38750ee874de · outbound

This paper cites an unresolved cited work.

Agent Safety Should Be a Runtime Contract Unresolved cited work

Reference 17

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.466753Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.466753Z digest=sha256:6cd1cebfe9a7313d42381bf11da3465e10c17d49e4497c1f8471562d07879ea0

Observation b6a9c064-3417-4b40-b15a-311292aeb1a0 · outbound

This paper cites Jailbreaking Black Box Large Language Models in Twenty Queries.

Agent Safety Should Be a Runtime Contract Jailbreaking Black Box Large Language Models in Twenty Queries

Reference 18

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.473161Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.473161Z digest=sha256:2b5f1165aefcafb9e672b70f859381771619c2e9fa4fe1c3a0381ed2aea7678e

Observation 4ec9e053-9424-4206-bed4-98ab0f0f5e11 · outbound

This paper cites an unresolved cited work.

Agent Safety Should Be a Runtime Contract Unresolved cited work

Reference 19

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.479129Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.479129Z digest=sha256:4c16f4c86482ca2035fde79bbad151999f3bca0e37f49700b7cacfb590092a41

Observation 08bb434a-f3da-44eb-94fb-c902ae71fedd · outbound

This paper cites Moffatt v.

Agent Safety Should Be a Runtime Contract Moffatt v

Reference 20

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.484569Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.484569Z digest=sha256:50a07bff3d1e071b0adf83629109eb18bf70db5b2d4d63102fa01ae46eb2b644

Observation ad939f0b-0446-4f08-9c5f-1c35dfccd802 · outbound

This paper cites an unresolved cited work.

Agent Safety Should Be a Runtime Contract Unresolved cited work

Reference 21

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.494790Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.494790Z digest=sha256:9444ea4b39bcca0849f3cd321a2486237695bcfb1ef2833f46dcfaf6c5e4d466

Observation f961d146-13a8-491a-83c8-3540e91039c7 · outbound

This paper cites an unresolved cited work.

Agent Safety Should Be a Runtime Contract Unresolved cited work

Reference 22

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.499446Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.499446Z digest=sha256:5b8d7dbc11bfa0b1c9a94eb57546a06ccd7a976ae40f11511ddbf5539d017492

Observation 05dc8549-d19e-4ed1-9072-2b303d85447b · outbound

This paper cites claude-code-log: Convert Claude Code JSONL transcripts to HTML, 2024.

Agent Safety Should Be a Runtime Contract claude-code-log: Convert Claude Code JSONL transcripts to HTML, 2024

Reference 23

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.504436Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.504436Z digest=sha256:c425557ab119a8d37141eb11728c494c420caaf21b99698b9fcaba58b5d6896d

Observation d51ffffa-a883-4d2b-a155-8a4ffc41182a · outbound

This paper cites Towards Guaranteed Safe AI: A Framework for Ensuring Robust and Reliable AI Systems.

Agent Safety Should Be a Runtime Contract Towards Guaranteed Safe AI: A Framework for Ensuring Robust and Reliable AI Systems

Reference 24

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.509712Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.509712Z digest=sha256:560c43457e034ad9ccf5531da8677ce25f2b744419e5e9f46ec9c612b4bc2e78

Observation 0c660fe8-5e3e-47a7-8178-33a246c3b50e · outbound

This paper cites Elhage, T.

Agent Safety Should Be a Runtime Contract Elhage, T

Reference 25

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.515708Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.515708Z digest=sha256:e6237818a62f7dd02d9e81d5b3415f0d151413c4954e6bf0b1b7ff84f6ecfb5d

Observation 83e6379b-65cb-4997-8986-e6e5dc45ae11 · outbound

This paper cites Regulation (EU) 2024/1689 laying down harmonised rules on artificial intelligence (AI act).Official Journal of the European Union, 2024.

Agent Safety Should Be a Runtime Contract Regulation (EU) 2024/1689 laying down harmonised rules on artificial intelligence (AI act).Official Journal of the European Union, 2024

Reference 26

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.520878Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.520878Z digest=sha256:4a7b2e7db0108a8ee22b81cac1677e7af515aaeb1f5ab40b00bd9a040e469315

Observation 4d7a03f7-3091-4909-838b-e3d39c6e2769 · outbound

This paper cites Fowler and M.

Agent Safety Should Be a Runtime Contract Fowler and M

Reference 27

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.526473Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.526473Z digest=sha256:f8ff6fb7057f2a7225fe2d3b71f5f16cf5db6ed782eb0655e7a4f328be246c45

Observation 9e6bc882-05c8-4323-adfd-b4f0a136ab52 · outbound

This paper cites Scaling laws: Eugene V olokh on navigating libel and liability in the AI age, 2024.

Agent Safety Should Be a Runtime Contract Scaling laws: Eugene V olokh on navigating libel and liability in the AI age, 2024

Reference 28

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.531892Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.531892Z digest=sha256:f0b290ed010827615bd66e18d64fe7eb6b5abdc69bf018d6447b47f197ac54ac

Observation a16c3b89-9865-418b-8295-37489f0f862c · outbound

This paper cites Red Teaming Language Models to Reduce Harms: Methods, Scaling Behaviors, and Lessons Learned.

Agent Safety Should Be a Runtime Contract Red Teaming Language Models to Reduce Harms: Methods, Scaling Behaviors, and Lessons Learned

Reference 29

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.537474Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.537474Z digest=sha256:ecba9b171146c21f8a96abb6634127143a9776b800e7654068b0821688842eb6

Observation 2797d8df-1484-44e1-9f12-226e3612eb6a · outbound

This paper cites an unresolved cited work.

Agent Safety Should Be a Runtime Contract Unresolved cited work

Reference 30

Resolution
unresolved
raw_fallback, observed 2026-08-15T14:19:16.492099Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.543513Z digest=sha256:3a7d97af051147e5ebed3c361af3e106df7edc0b9a1f1fd4e1879bb48071288f

Observation 607e55b2-5fe5-4b50-8bbb-7a9a7539fa38 · outbound

This paper cites Gauthier.

Agent Safety Should Be a Runtime Contract Gauthier

Reference 31

Resolution
verified fuzzy
raw_fallback, observed 2026-08-15T14:19:16.462278Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.548339Z digest=sha256:0dbf663294cc1a1b60c7e215a8b6a9034e0adbde852b0b37e53e98468727f014

Observation d6b4f02f-af83-4d67-8492-be17a4662931 · outbound

This paper cites Amazon experiences AWS outage believed to be caused by AI tools, 2026.

Agent Safety Should Be a Runtime Contract Amazon experiences AWS outage believed to be caused by AI tools, 2026

Reference 32

Resolution
verified exact
raw_fallback, observed 2026-08-15T14:19:14.814918Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.553062Z digest=sha256:2792f6aab489b0c700ac11432d8ecf473f9f22ff622d70e46e3a356607c5b041

Observation bd468bd4-2e91-4674-b40b-168f3a4f5487 · outbound

This paper cites Tracking GitHub Copilot’s sessions, 2024.

Agent Safety Should Be a Runtime Contract Tracking GitHub Copilot’s sessions, 2024

Reference 33

Resolution
verified fuzzy
raw_fallback, observed 2026-08-15T14:19:16.427259Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.558277Z digest=sha256:cc702684dfc8860a8bf2386c07d5fb3e89d60f8daeff81352d915cd69800b161

Observation f9213660-11bc-473c-842a-b5ed8c9cd328 · outbound

This paper cites Alignment faking in large language models.

Agent Safety Should Be a Runtime Contract Alignment faking in large language models

Reference 34

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.563860Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.563860Z digest=sha256:2ced78a0ff645c99f5bf3c0d6d071fade4266662c7d4c6cffc5960a613485bb9

Observation b1a7d303-c2bb-4073-a3a2-f77e218946b2 · outbound

This paper cites an unresolved cited work.

Agent Safety Should Be a Runtime Contract Unresolved cited work

Reference 35

Resolution
unresolved
raw_fallback, observed 2026-08-15T14:19:16.397836Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.570293Z digest=sha256:945576d9180c93a05da620d276c224ea5d6c999e4eca129acf27a54cf17397f4

Observation 826c1b20-29b6-4887-9921-9c3d7aa06bd3 · outbound

This paper cites Cursor IDE support hallucinates lockout policy, causes user cancellations, 2025.

Agent Safety Should Be a Runtime Contract Cursor IDE support hallucinates lockout policy, causes user cancellations, 2025

Reference 36

Resolution
verified fuzzy
raw_fallback, observed 2026-08-15T14:19:16.372883Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.578110Z digest=sha256:35d3e38b25c09ffa24ca58cd10029e130fae5e764b7f001f6e3cd176f933f770

Observation 14b76236-b1bf-4bae-9d1b-08978a960d97 · outbound

This paper cites an unresolved cited work.

Agent Safety Should Be a Runtime Contract Unresolved cited work

Reference 37

Resolution
unresolved
raw_fallback, observed 2026-08-15T14:19:16.343279Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.585846Z digest=sha256:0bfa5084bae12ae90e578251734af9da2c698deced67bc2a35064b375d008f5b

Observation d482a7e1-fd71-48f6-8c0a-a2f9f8a1c9c4 · outbound

This paper cites an unresolved cited work.

Agent Safety Should Be a Runtime Contract Unresolved cited work

Reference 38

Resolution
unresolved
raw_fallback, observed 2026-08-15T14:19:16.316434Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.594864Z digest=sha256:e8e18b5d9c2c4ccb09322e13d57cd4730945ff00ef0afb30c0ad52043af67491

Observation 59200826-9f54-46be-a9ea-6eeab349cb2d · outbound

This paper cites Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training.

Agent Safety Should Be a Runtime Contract Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training

Reference 39

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.601493Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.601493Z digest=sha256:203100e0abe58a662d707ea6066563e8665bd68dc03a62a5c1450efcb1457b19

Observation ba981562-dabe-4ac1-9a91-9949be5dced0 · outbound

This paper cites Humble and D.

Agent Safety Should Be a Runtime Contract Humble and D

Reference 40

Resolution
verified fuzzy
raw_fallback, observed 2026-08-15T14:19:16.291087Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.606923Z digest=sha256:9d310d21a77ceba96e87838bcf11caece1129b8cdcaa790d7f34c286e142c28a

Observation 647eb7ab-67a5-4c28-809a-7929790e0c19 · outbound

This paper cites Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversations.

Agent Safety Should Be a Runtime Contract Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversations

Reference 41

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.614533Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.614533Z digest=sha256:1ea4240d391bc3d3090ae7b39cee401a2e7272d8dad191b4b9d1924b78658f15

Observation 6db12924-2562-4728-8e98-9dc4f10e8115 · outbound

This paper cites an unresolved cited work.

Agent Safety Should Be a Runtime Contract Unresolved cited work

Reference 42

Resolution
unresolved
raw_fallback, observed 2026-08-15T14:19:16.263181Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.620535Z digest=sha256:ef770ce411337b392970be989291e9c4c119fc686b023920cd7744912eea8980

Observation f9d9f766-9444-4f59-8b91-85b9e3587c38 · outbound

This paper cites Kent and M.

Agent Safety Should Be a Runtime Contract Kent and M

Reference 43

Resolution
verified fuzzy
raw_fallback, observed 2026-08-15T14:19:16.230946Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.631244Z digest=sha256:a1aabb6f02c9bec46b360b0baef3356fe38f0aa9365161aef0ede1b0e7223620

Observation 3b6aada0-f6bc-4719-a952-fb3fec25c4c2 · outbound

This paper cites an unresolved cited work.

Agent Safety Should Be a Runtime Contract Unresolved cited work

Reference 44

Resolution
unresolved
raw_fallback, observed 2026-08-15T14:19:16.177973Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.637494Z digest=sha256:1098533a91d8d13dc247bb173a807253dd5be9ac0cb22c5affb4512695793e78

Observation fab181d6-dc20-4357-bba6-07edc77db1b3 · outbound

This paper cites Metr’s observations of reward hacking in recent frontier models, 2025.

Agent Safety Should Be a Runtime Contract Metr’s observations of reward hacking in recent frontier models, 2025

Reference 45

Resolution
verified fuzzy
raw_fallback, observed 2026-08-15T14:19:16.144969Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.645514Z digest=sha256:700fe7a7525dd42a873f7af72343104d2638847494c8066363214f5e86857b46

Observation 008dbdb3-f542-4897-a3a9-4530d36641f3 · outbound

This paper cites Krakovna, J.

Agent Safety Should Be a Runtime Contract Krakovna, J

Reference 46

Resolution
verified fuzzy
raw_fallback, observed 2026-08-15T14:19:16.113291Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.653020Z digest=sha256:03b2dbc349745c558d25bda30e0f1cf308d40c7fe5f489cf5d795869aa4e31e2

Observation f00b8d74-2042-425e-ba3d-efcbbd881f5f · outbound

This paper cites Kupferman and M.

Agent Safety Should Be a Runtime Contract Kupferman and M

Reference 47

Resolution
verified fuzzy
raw_fallback, observed 2026-08-15T14:19:16.084455Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.659030Z digest=sha256:c50363e0ed50b122524ad2fb249035c3f6269ecdf1e623add7029cfecd707827

Observation 146a0071-30d7-4c12-81a9-771dac776786 · outbound

This paper cites an unresolved cited work.

Agent Safety Should Be a Runtime Contract Unresolved cited work

Reference 48

Resolution
unresolved
raw_fallback, observed 2026-08-15T14:19:16.050980Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.664734Z digest=sha256:aeccb5a60c6420c245dc447fa94c2ad6a36289c1a4b7616ae76c5c1d7c6fac47

Observation 2d0826db-f657-4363-ad14-aface78d1643 · outbound

This paper cites an unresolved cited work.

Agent Safety Should Be a Runtime Contract Unresolved cited work

Reference 49

Resolution
unresolved
raw_fallback, observed 2026-08-15T14:19:16.021805Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.671476Z digest=sha256:c295682bc0ad68d7e3eecbf5291387b12173add7e7274e4636ef6cf6114dd930

Observation 6a2e64a5-86ef-4f2d-823b-9e8389c5da08 · outbound

This paper cites RLAIF vs. RLHF: Scaling Reinforcement Learning from Human Feedback with AI Feedback.

Agent Safety Should Be a Runtime Contract RLAIF vs. RLHF: Scaling Reinforcement Learning from Human Feedback with AI Feedback

Reference 50

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.678318Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.678318Z digest=sha256:ae2a17c265c9da542c883ecb6a7e258f4613d61466aaa358dc803a7373ae5556

Observation 2c36c004-fd47-42a3-9f2d-1ac7237d8172 · outbound

This paper cites an unresolved cited work.

Agent Safety Should Be a Runtime Contract Unresolved cited work

Reference 51

Resolution
unresolved
raw_fallback, observed 2026-08-15T14:19:15.994487Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.684448Z digest=sha256:6ee3b8ed507f36141c38f2e07195d9f4ce3d62ac141e89738f504276517f5fa1

Observation a22eb722-6441-4efd-acf2-8be14d14da10 · outbound

This paper cites Leucker and C.

Agent Safety Should Be a Runtime Contract Leucker and C

Reference 52

Resolution
verified fuzzy
raw_fallback, observed 2026-08-15T14:19:15.957381Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.689649Z digest=sha256:b434a4ce79a6fb70727fffb299f08739e3cb0a185eb333cd7ba45a1963b22296

Observation 6d9f6346-046b-45a8-a25c-ebf7d260740f · outbound

This paper cites an unresolved cited work.

Agent Safety Should Be a Runtime Contract Unresolved cited work

Reference 53

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.694535Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.694535Z digest=sha256:64bfd56895ba588cd33048d2ccfd76a9c14c5cf5afda0654df94e78418b1e9ae

Observation 7c8d4f2d-b1e1-4e22-9b41-431b7bc9b40d · outbound

This paper cites an unresolved cited work.

Agent Safety Should Be a Runtime Contract Unresolved cited work

Reference 54

Resolution
unresolved
raw_fallback, observed 2026-08-15T14:19:15.913924Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.700011Z digest=sha256:f013b2f410917cedc7106708360cffb11fbd178f41011a4769e54155679af1ab

Observation 1c901f69-299f-43a1-97b0-3555eff0071d · outbound

This paper cites HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal.

Agent Safety Should Be a Runtime Contract HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal

Reference 55

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.706357Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.706357Z digest=sha256:9b0adccf84aa2af03923f17aedbcc9d91733eef040ea6f82b75792cf717a7e2f

Observation b0096995-792c-4392-af30-e72f00db3df4 · outbound

This paper cites McGregor.

Agent Safety Should Be a Runtime Contract McGregor

Reference 56

Resolution
verified fuzzy
raw_fallback, observed 2026-08-15T14:19:15.892904Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.713244Z digest=sha256:27230998c62b86320adfa481a060d726c90a27e731776007908f52370e6d66a1

Observation d41f58a9-7938-4021-8c95-4fd7b7f09d03 · outbound

This paper cites Meinke, B.

Agent Safety Should Be a Runtime Contract Meinke, B

Reference 57

Resolution
verified fuzzy
raw_fallback, observed 2026-08-15T14:19:15.870090Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.718160Z digest=sha256:3be89016196731cda4c763f7794610656d800c8fb76b9dc4e471ecfabf1232ca

Observation e4d2ec69-0ede-4330-8af6-2583379ba992 · outbound

This paper cites an unresolved cited work.

Agent Safety Should Be a Runtime Contract Unresolved cited work

Reference 58

Resolution
unresolved
raw_fallback, observed 2026-08-15T14:19:15.840822Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.723693Z digest=sha256:c0bd3eca2ea333cbe4f1ef1b7877783d5a735ded2be9da3b2a44864b4fb69a32

Observation 78c5b4f5-9f6a-4df2-aff2-6f1306d22cd2 · outbound

This paper cites Llama guard 3.Meta AI Technical Report, 2024.https://www.llama.com/docs/model-car ds-and-prompt-formats/llama-guard-3/.

Agent Safety Should Be a Runtime Contract Llama guard 3.Meta AI Technical Report, 2024.https://www.llama.com/docs/model-car ds-and-prompt-formats/llama-guard-3/

Reference 59

Resolution
verified fuzzy
raw_fallback, observed 2026-08-15T14:19:15.814781Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.731458Z digest=sha256:4651297e6c83dda4e78d8e5b96df970fd4322a05ae547eea3e4309a60a52e541

Observation b5e1172e-b5c9-459c-84df-fe8b695521f0 · outbound

This paper cites Nanda, L.

Agent Safety Should Be a Runtime Contract Nanda, L

Reference 60

Resolution
verified fuzzy
raw_fallback, observed 2026-08-15T14:19:15.796473Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.736922Z digest=sha256:ffeaa3106a9a2cd4b7be810a6a6a990cfb9af1453271ccd49804c647dfe36a4f

Observation ff71ffe7-4365-46e4-afd1-bd46b8cd49d6 · outbound

This paper cites Framework for improving critical infrastructure cyberse- curity, version 1.1.

Agent Safety Should Be a Runtime Contract Framework for improving critical infrastructure cyberse- curity, version 1.1

Reference 61

Resolution
verified fuzzy
raw_fallback, observed 2026-08-15T14:19:15.776878Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.742700Z digest=sha256:255e1d060061bd57fb2fa51b6b85d4148c2c05dde9383c810a3f6886497c22fc

Observation 0d828cbd-d76b-44af-91a5-741cdf015055 · outbound

This paper cites Artificial intelligence risk management framework (AI RMF 1.0).

Agent Safety Should Be a Runtime Contract Artificial intelligence risk management framework (AI RMF 1.0)

Reference 62

Resolution
verified fuzzy
raw_fallback, observed 2026-08-15T14:19:15.754172Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.747660Z digest=sha256:699de42906ce3ec0e8b2f12a22483f926d89c27e839be9c83cc33824b8c1da3d

Observation 4ed172c8-146c-4911-a464-f720be08d607 · outbound

This paper cites ChatGPT wrongly accuses law professor of sexual assault, 2023.

Agent Safety Should Be a Runtime Contract ChatGPT wrongly accuses law professor of sexual assault, 2023

Reference 63

Resolution
verified fuzzy
raw_fallback, observed 2026-08-15T14:19:15.730450Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.753519Z digest=sha256:02f447301f493e83d046b63274790b6aebed49cffe26bf1c664032fc81a2e240

Observation 8d9206c6-df64-428c-bbd8-cc9f7e5d501e · outbound

This paper cites an unresolved cited work.

Agent Safety Should Be a Runtime Contract Unresolved cited work

Reference 64

Resolution
unresolved
raw_fallback, observed 2026-08-15T14:19:15.709445Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.760190Z digest=sha256:df8367e389b0cf7f27c8cbdb9a84736025e49a343299e1282053b0fd363160ce

Observation 1acd86e8-ab82-4ed5-a9a7-6bc3aef8e2ba · outbound

This paper cites GPT-4 Technical Report.

Agent Safety Should Be a Runtime Contract GPT-4 Technical Report

Reference 65

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.765269Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.765269Z digest=sha256:65ef71994cc1003067c41c4015ccd2e5eba2a85404596a6722d566895f0d2f45

Observation 49a61a66-d008-44a9-9961-405dfe183e12 · outbound

This paper cites Codex CLI features, 2025.https://developers.openai.com/codex/cli/features.

Agent Safety Should Be a Runtime Contract Codex CLI features, 2025.https://developers.openai.com/codex/cli/features

Reference 66

Resolution
verified fuzzy
raw_fallback, observed 2026-08-15T14:19:15.689475Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.771262Z digest=sha256:650faf7083e5b13712a3866882eac66534a267703db6f882d07a80c9029a71c2

Observation 128b59bd-d578-4fe6-bc84-91a9e3de3656 · outbound

This paper cites SWE-Bench Benchmark Evaluation, 2024.

Agent Safety Should Be a Runtime Contract SWE-Bench Benchmark Evaluation, 2024

Reference 67

Resolution
verified fuzzy
raw_fallback, observed 2026-08-15T14:19:15.666753Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.777797Z digest=sha256:c0b52e805103d10607f340c1cdd0e8088de05e916ed9105ad9eb814aa92e9559

Observation e58c50ce-d6f1-4293-a784-04d8aab235f8 · outbound

This paper cites Ouyang, J.

Agent Safety Should Be a Runtime Contract Ouyang, J

Reference 68

Resolution
verified fuzzy
raw_fallback, observed 2026-08-15T14:19:15.647397Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.785532Z digest=sha256:d137d1798ee18fb5d2a8fbc8b2ba4f7d96ccd8c4bdc296854d370c39a0b3cc75

Observation 083c23b3-45e0-4a42-90cf-5814352e8c44 · outbound

This paper cites Discovering Language Model Behaviors with Model-Written Evaluations.

Agent Safety Should Be a Runtime Contract Discovering Language Model Behaviors with Model-Written Evaluations

Reference 69

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.791329Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.791329Z digest=sha256:e9555b85a6af1db6122292e176b3239d3b23608c9ff16b566d6de13d933f5824

Observation f04bb7fd-7319-4ced-a27d-e81afd1c3207 · outbound

This paper cites an unresolved cited work.

Agent Safety Should Be a Runtime Contract Unresolved cited work

Reference 70

Resolution
verified exact
raw_fallback, observed 2026-08-15T14:19:14.526976Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.798151Z digest=sha256:7a8d7398b5adfc30b676e39687dbdb3834d13f2c381d82ce850c843e2f545d94

Observation e0c0f51d-dae9-44c8-af6a-d19a560c3abd · outbound

This paper cites Pnueli and R.

Agent Safety Should Be a Runtime Contract Pnueli and R

Reference 71

Resolution
verified fuzzy
raw_fallback, observed 2026-08-15T14:19:15.614741Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.804275Z digest=sha256:9e00b6ce936c927ee574972fd4772f1e687c55165286aaaec001866610b96034

Observation 02a5a708-47a6-4117-b01b-83619eb5d08b · outbound

This paper cites Provos, M.

Agent Safety Should Be a Runtime Contract Provos, M

Reference 72

Resolution
verified fuzzy
raw_fallback, observed 2026-08-15T14:19:15.590991Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.811738Z digest=sha256:ad18a09b48c315a8e44ac237ddf9bf59a56d3a44055ec15e52753a78c04e096b

Observation 250cfaab-63c8-401b-bd14-76ee2e991516 · outbound

This paper cites Fine-tuning Aligned Language Models Compromises Safety, Even When Users Do Not Intend To!.

Agent Safety Should Be a Runtime Contract Fine-tuning Aligned Language Models Compromises Safety, Even When Users Do Not Intend To!

Reference 73

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.818100Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.818100Z digest=sha256:add7602464500a65a2a2aa83ead5968c69da4d648ad97fe7a2768f1327b3f617

Observation 8459074b-1432-4951-a16b-eb91c2fa895f · outbound

This paper cites Rafailov, A.

Agent Safety Should Be a Runtime Contract Rafailov, A

Reference 74

Resolution
verified fuzzy
raw_fallback, observed 2026-08-15T14:19:15.567145Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.824322Z digest=sha256:e1f772fdbc6d0043b600e77fe559c8a4e7e79fae1a58d53af96ad4946baf5dd7

Observation 03571801-fe91-4608-948f-f982897c8e80 · outbound

This paper cites NeMo Guardrails: A Toolkit for Controllable and Safe LLM Applications with Programmable Rails.

Agent Safety Should Be a Runtime Contract NeMo Guardrails: A Toolkit for Controllable and Safe LLM Applications with Programmable Rails

Reference 75

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.829661Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.829661Z digest=sha256:f596431f33777e378756021653d41d5848e7507969f706cfd91c1cc50e0cb6fa

Observation 25e91600-2c4c-42de-9ba0-b71b0d95558b · outbound

This paper cites Reddy and A.

Agent Safety Should Be a Runtime Contract Reddy and A

Reference 76

Resolution
verified fuzzy
raw_fallback, observed 2026-08-15T14:19:15.546211Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.836434Z digest=sha256:c0ffd3343f56e953b7df91d821c5a7f5e211bee199f028dafd94cab084e34a20

Observation af9fe6f4-5e54-40de-b42b-acd7e9ebb66e · outbound

This paper cites an unresolved cited work.

Agent Safety Should Be a Runtime Contract Unresolved cited work

Reference 77

Resolution
unresolved
raw_fallback, observed 2026-08-15T14:19:15.520745Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.846572Z digest=sha256:31f64b66659593a51554e5d0b0fb8196152f944d3fb885aa4644a1f5f993a90d

Observation 62150ed3-0f2a-49da-b073-1f4de9abcef1 · outbound

This paper cites Robinson.

Agent Safety Should Be a Runtime Contract Robinson

Reference 78

Resolution
verified fuzzy
raw_fallback, observed 2026-08-15T14:19:15.494107Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.853115Z digest=sha256:fd6fbe41801da54acfc05338cff6bdd76983bcf3c5c801395d95c01edde69843

Observation f1638fce-bec2-4e81-b6e4-5734f4760968 · outbound

This paper cites Identifying the Risks of LM Agents with an LM-Emulated Sandbox.

Agent Safety Should Be a Runtime Contract Identifying the Risks of LM Agents with an LM-Emulated Sandbox

Reference 79

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.859825Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.859825Z digest=sha256:789b0eab8575b922b7bb9f349ab6df41b5848832cbf5e8e2bb854a3a50f06678

Observation 779109d4-8284-408a-aebe-1b518aa79e36 · outbound

This paper cites an unresolved cited work.

Agent Safety Should Be a Runtime Contract Unresolved cited work

Reference 80

Resolution
unresolved
raw_fallback, observed 2026-08-15T14:19:15.472021Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.867788Z digest=sha256:aa2dd55a27d37faa667d379ce3b928b664f43ad8f8693acbf4c8ce48a4bd5987

Observation b779c20c-ec17-40f6-9b29-dd04f113f947 · outbound

This paper cites an unresolved cited work.

Agent Safety Should Be a Runtime Contract Unresolved cited work

Reference 81

Resolution
unresolved
raw_fallback, observed 2026-08-15T14:19:15.438541Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.875086Z digest=sha256:ba4c9f826e73277203f604fcc2bbe67ed63f6a2f38ea866156bf69ee7415f5ce

Observation 0bcd76f6-e3ac-4a7b-a8e9-c64f2a19270e · outbound

This paper cites Towards Understanding Sycophancy in Language Models.

Agent Safety Should Be a Runtime Contract Towards Understanding Sycophancy in Language Models

Reference 82

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.880401Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.880401Z digest=sha256:1854d4246138ecd9d766cf4d14803a35923da40172f76b7133d2d2604f36644f

Observation 574972e0-1c78-4261-a972-7631964dcc8b · outbound

This paper cites Sharwood.

Agent Safety Should Be a Runtime Contract Sharwood

Reference 83

Resolution
verified fuzzy
raw_fallback, observed 2026-08-15T14:19:15.413474Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.886530Z digest=sha256:eca838457777a5989ea136e4d07c7385ee8fed2508b44f4d87def5bcd765cd30

Observation eb9abfb2-6f1c-47d1-abad-2f1b880472d7 · outbound

This paper cites Shavit et al.

Agent Safety Should Be a Runtime Contract Shavit et al

Reference 84

Resolution
verified fuzzy
raw_fallback, observed 2026-08-15T14:19:15.371441Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.899635Z digest=sha256:4e8265150d53f4fa312a70ad982d4658df3e32ae76c29abf14abb8b57d36ebc9

Observation 0013899b-85a3-4ce1-ac6e-4d2af1b237da · outbound

This paper cites do anything now.

Agent Safety Should Be a Runtime Contract do anything now

Reference 85

Resolution
verified fuzzy
raw_fallback, observed 2026-08-15T14:19:15.350995Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.904688Z digest=sha256:b09b405f60f3b72eaf07ff55980d84dc5c07d2ff0f413160133e2e3c3a865d78

Observation caea7364-5a45-4cbe-8909-03661c91b86c · outbound

This paper cites Skalse, N.

Agent Safety Should Be a Runtime Contract Skalse, N

Reference 86

Resolution
verified fuzzy
raw_fallback, observed 2026-08-15T14:19:15.323017Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.910758Z digest=sha256:cb4f6834ce0844e9684f38c15e248db964f1d5655930fe514901a6556618ef7e

Observation 3c364802-eec7-4b6d-985d-7162aee549c0 · outbound

This paper cites Stiennon, L.

Agent Safety Should Be a Runtime Contract Stiennon, L

Reference 87

Resolution
verified fuzzy
raw_fallback, observed 2026-08-15T14:19:15.287454Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.916421Z digest=sha256:f739f0329a2e2b135cdc85cba782aa80d226efd7390951952129bed9ce361d02

Observation 8e78c707-bec2-4671-868a-d8259b85f6fc · outbound

This paper cites an unresolved cited work.

Agent Safety Should Be a Runtime Contract Unresolved cited work

Reference 88

Resolution
metadata mismatch
raw_fallback, observed 2026-08-15T14:19:14.334812Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.924164Z digest=sha256:60deffbc53874c4fc0ce3cf8cea1e241a17dcf7ee1c35611a152e9149eaf8e9f

Observation d9579374-d049-48f1-86d3-7e38a3f02e13 · outbound

This paper cites Templeton, T.

Agent Safety Should Be a Runtime Contract Templeton, T

Reference 89

Resolution
verified fuzzy
raw_fallback, observed 2026-08-15T14:19:15.257243Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.931607Z digest=sha256:576e6b577f042f44eec93d18bd59ab7f80b61b0c47977eb8010017ce4a43a624

Observation bc766391-3757-4d2f-8079-0d73cf1b2f32 · outbound

This paper cites an unresolved cited work.

Agent Safety Should Be a Runtime Contract Unresolved cited work

Reference 90

Resolution
unresolved
raw_fallback, observed 2026-08-15T14:19:15.227550Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.937890Z digest=sha256:393cb39079d4c0bfe1889872fa850a0442fd09ebfd7fea13c16d93e234c71c4e

Observation 4b444593-a4db-43fa-8e7e-af01a62f347d · outbound

This paper cites Recent frontier models are reward hacking.

Agent Safety Should Be a Runtime Contract Recent frontier models are reward hacking

Reference 91

Resolution
verified fuzzy
raw_fallback, observed 2026-08-15T14:19:15.195866Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.943645Z digest=sha256:8215b11454a20f5ab5aaf4eec651bda0a13c5913fe9aee62d4419aac227e2d89

Observation d72e8010-6c07-41a6-af8b-07ab0469993a · outbound

This paper cites D$^2$LV: A Data-Driven and Local-Verification Approach for Image Copy Detection.

Agent Safety Should Be a Runtime Contract D$^2$LV: A Data-Driven and Local-Verification Approach for Image Copy Detection

Reference 92

Resolution
verified exact
local_arxiv, observed 2026-08-15T14:19:14.073142Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.949193Z digest=sha256:ac1119d0779596ae31094ddcf1c259b18e0008a6cf133b09c57bf2141cf9761a

Observation 7c35d54b-82b3-4d5d-85f2-ddb367d14286 · outbound

This paper cites Are "Solved Issues" in SWE-bench Really Solved Correctly? An Empirical Study.

Agent Safety Should Be a Runtime Contract Are "Solved Issues" in SWE-bench Really Solved Correctly? An Empirical Study

Reference 93

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.956774Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.956774Z digest=sha256:1eb37d96cb9b4d3ca4e28463520c0fd4c387915bfd423e473c308dafe3f28712

Observation a7d84a3b-a8dc-4882-a567-9ab7fd677461 · outbound

This paper cites an unresolved cited work.

Agent Safety Should Be a Runtime Contract Unresolved cited work

Reference 94

Resolution
unresolved
raw_fallback, observed 2026-08-15T14:19:15.160226Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.965986Z digest=sha256:46b8e7b17be6d61e4cd4225cff5486da847a895043d6743868cc881a8a7f471f

Observation 7e2e9e9a-7fe0-492a-a0a9-d4124f888d87 · outbound

This paper cites an unresolved cited work.

Agent Safety Should Be a Runtime Contract Unresolved cited work

Reference 95

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.971491Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.971491Z digest=sha256:4f87d49f8c9d973b7f8ae8d9a33ba330c7c4590ab91f0313c717e2233eb7e0c4

Observation e726e422-6de1-4807-a43c-ae8942b85159 · outbound

This paper cites Fundamental Limitations of Alignment in Large Language Models.

Agent Safety Should Be a Runtime Contract Fundamental Limitations of Alignment in Large Language Models

Reference 96

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.977631Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.977631Z digest=sha256:e73aba3ae74cfdc5c494f96a7484579b15f6ccdb31f1810d51549c4edc74f850

Observation 933e9d9a-13e5-49af-ab4a-1157bbacc3ec · outbound

This paper cites Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models.

Agent Safety Should Be a Runtime Contract Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models

Reference 97

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:13.984292Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:13.984292Z digest=sha256:ce8df7367369e47587ac4d719ba7aa5dc36578d419bb0e8dded9bb048b67107b

Observation ac5ecb8a-d6f5-4ed7-8d00-2880bee74e46 · outbound

This paper cites an unresolved cited work.

Agent Safety Should Be a Runtime Contract Unresolved cited work

Reference 98

Resolution
unresolved
raw_fallback, observed 2026-08-15T14:19:15.117217Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.992674Z digest=sha256:6431867915e057edc2acb4e273a0f28c589ea2cb03e3e86b75913928c1be45d6

Observation cf794151-8f81-49e3-a304-ae80f65e6efb · outbound

This paper cites hallucinates.

Agent Safety Should Be a Runtime Contract hallucinates

Reference 99

Resolution
verified fuzzy
raw_fallback, observed 2026-08-15T14:19:15.094474Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-18T06:34:40.430872+00:00.

source=pdf_text observed=2026-08-15T14:19:13.999238Z digest=sha256:2df00f352ddebab0b3ee8c2cd855d1b42d0a29569c367608469f963f7fe2663c

Observation e2b186c1-b8e7-4440-af39-d8191b3b9363 · outbound

This paper cites Establishing Best Practices for Building Rigorous Agentic Benchmarks.

Agent Safety Should Be a Runtime Contract Establishing Best Practices for Building Rigorous Agentic Benchmarks

Reference 100

Resolution
unresolved
no resolver link, observed 2026-08-15T14:19:14.004821Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-15T14:19:14.004821Z digest=sha256:4153fb890e90582dc3c724687ca7cdbd9db4e0a52812c7eb2eb7eb1d35a5d29e

Pith citing papers

No inbound Pith citation observations are available.