Pith. sign in

Paper Citation Record · LEDGER

Phantom Guardrails: When Self-Improving Agent Harnesses Fix Failures That Never Happened

As of 18 August 2026, this Paper Citation Record lists 28 of 28 outbound references and 0 inbound Pith citation observations for arXiv:2607.13083.

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

pith.paper-citation-record.v1
2607.13083 v1

Coverage vector

measured 28 of 28 reference resolution

Typed states for the displayed outbound observations.

Source: paper_references, paper_reference_links, observed 2026-08-02T07:01:23.147046Z

measured 28 of 28 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

28 of 28 outbound references displayed

  • verified exact0
  • verified fuzzy0
  • unresolved27
  • parse uncertain0
  • malformed identifier1
  • metadata mismatch0

External citation measurements

No source-named external measurement is stored.

Outbound references

Observation af5421f9-5c79-4bd3-909e-28d7a610d778 · outbound

This paper cites OR-Bench: An Over-Refusal Benchmark for Large Language Models.

Phantom Guardrails: When Self-Improving Agent Harnesses Fix Failures That Never Happened OR-Bench: An Over-Refusal Benchmark for Large Language Models

Reference 1

Resolution
unresolved
no resolver link, observed 2026-08-02T07:01:19.315752Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-02T07:01:19.315752Z digest=sha256:09aff96c934bd66d954e0ee3d9d1e5c07c4b863722f145264e8aaeb40ade4e1f

Observation f7ef4e9a-452b-4073-a590-901d3f6eefe4 · outbound

This paper cites an unresolved cited work.

Phantom Guardrails: When Self-Improving Agent Harnesses Fix Failures That Never Happened Unresolved cited work

Reference 2

Resolution
unresolved
no resolver link, observed 2026-08-02T07:01:19.498726Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-02T07:01:19.498726Z digest=sha256:c8aa4c00c68772faf67b59e304465f30d54c162641aebfdc812220b2562c3055

Observation d4d35aed-bf20-45e4-8ac7-d0cb115867fd · outbound

This paper cites SCRIBE: Structured Mid-Level Supervision for Tool-Using Language Models.

Phantom Guardrails: When Self-Improving Agent Harnesses Fix Failures That Never Happened SCRIBE: Structured Mid-Level Supervision for Tool-Using Language Models

Reference 3

Resolution
unresolved
no resolver link, observed 2026-08-02T07:01:19.651690Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-02T07:01:19.651690Z digest=sha256:b49680cb81de7e950bc0cf63264d6a9d1ff633356bd666916d3eaa3999229600

Observation 4f1c9d35-4cd3-44ad-b3ff-bab706e418c4 · outbound

This paper cites an unresolved cited work.

Phantom Guardrails: When Self-Improving Agent Harnesses Fix Failures That Never Happened Unresolved cited work

Reference 4

Resolution
unresolved
no resolver link, observed 2026-08-02T07:01:19.768759Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-02T07:01:19.768759Z digest=sha256:d3e899ea82261d67c2b247ffac7a8c583027b4e98b10270335453b8e34a3791a

Observation 662be392-25cb-4fda-a729-5dbf7a3ac4d7 · outbound

This paper cites Meta-Harness: End-to-End Optimization of Model Harnesses.

Phantom Guardrails: When Self-Improving Agent Harnesses Fix Failures That Never Happened Meta-Harness: End-to-End Optimization of Model Harnesses

Reference 5

Resolution
unresolved
no resolver link, observed 2026-08-02T07:01:20.047229Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-02T07:01:20.047229Z digest=sha256:4a263af3d94a38fec9a11d55ec2dab6f520475d618a6691ab165bd1dc44472fb

Observation b2382074-deff-4c7a-a6d4-0d34431547a0 · outbound

This paper cites Analysing Chain of Thought Dynamics: Active Guidance or Unfaithful Post-hoc Rationalisation?.

Phantom Guardrails: When Self-Improving Agent Harnesses Fix Failures That Never Happened Analysing Chain of Thought Dynamics: Active Guidance or Unfaithful Post-hoc Rationalisation?

Reference 6

Resolution
unresolved
no resolver link, observed 2026-08-02T07:01:20.213268Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-02T07:01:20.213268Z digest=sha256:619ad0f72b22a5fd561555e649ee8f04fcf813dbe25bf5c117b713d07b1b4c07

Observation 0c42fcc9-aff9-47f8-83a0-0293f9a6fefa · outbound

This paper cites an unresolved cited work.

Phantom Guardrails: When Self-Improving Agent Harnesses Fix Failures That Never Happened Unresolved cited work

Reference 7

Resolution
unresolved
no resolver link, observed 2026-08-02T07:01:20.358148Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-02T07:01:20.358148Z digest=sha256:7538429f4c332b5cba545d654bf1c5835a4666666c0cd9b41fc244b0bb47f45b

Observation 3b8334a4-eacb-4c28-8943-1f898beebaf5 · outbound

This paper cites SafetyRepro: Configuration-Conditional Rank Instability on Alignment Benchmarks.

Phantom Guardrails: When Self-Improving Agent Harnesses Fix Failures That Never Happened SafetyRepro: Configuration-Conditional Rank Instability on Alignment Benchmarks

Reference 8

Resolution
unresolved
no resolver link, observed 2026-08-02T07:01:20.541311Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-02T07:01:20.541311Z digest=sha256:45596c9873fa21134d0bd192b6c0772ba9d216430b1002cc06243ceaad1ef02e

Observation f4a52139-6642-4601-b802-6722bdf0fd23 · outbound

This paper cites Agentic Harness Engineering: Observability-Driven Automatic Evolution of Coding-Agent Harnesses.

Phantom Guardrails: When Self-Improving Agent Harnesses Fix Failures That Never Happened Agentic Harness Engineering: Observability-Driven Automatic Evolution of Coding-Agent Harnesses

Reference 9

Resolution
unresolved
no resolver link, observed 2026-08-02T07:01:20.674740Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-02T07:01:20.674740Z digest=sha256:e26079d62ff2de64c5da097c1480cb13558ab4dc1d704f9238cb88ef456741a7

Observation e5520ea9-d240-4176-85c2-40c56431da6d · outbound

This paper cites Reflect-Guard: Enhancing LLM Safeguards against Adversarial Prompts via Logical Self-Reflection.

Phantom Guardrails: When Self-Improving Agent Harnesses Fix Failures That Never Happened Reflect-Guard: Enhancing LLM Safeguards against Adversarial Prompts via Logical Self-Reflection

Reference 10

Resolution
unresolved
no resolver link, observed 2026-08-02T07:01:20.817475Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-02T07:01:20.817475Z digest=sha256:45cc8398331348082bbb74c605f318333c9d09ba0ad3a569d1f7d0b57496c50a

Observation 83cb1264-bb3c-481a-95f8-6be96b296c21 · outbound

This paper cites You Cannot Fix What You Cannot Find! An Investigation of Fault Localization Bias in Benchmarking Automated Program Repair Systems.

Phantom Guardrails: When Self-Improving Agent Harnesses Fix Failures That Never Happened You Cannot Fix What You Cannot Find! An Investigation of Fault Localization Bias in Benchmarking Automated Program Repair Systems

Reference 11

Resolution
unresolved
no resolver link, observed 2026-08-02T07:01:20.951539Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-02T07:01:20.951539Z digest=sha256:1dc4f8005d4e5f9e6f1cdac22e84bc56f0826715d7b42ee96bc5cf4fdc9b9744

Observation 44c67161-c09e-403a-8f6a-256350610aac · outbound

This paper cites an unresolved cited work.

Phantom Guardrails: When Self-Improving Agent Harnesses Fix Failures That Never Happened Unresolved cited work

Reference 12

Resolution
unresolved
no resolver link, observed 2026-08-02T07:01:21.021689Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-02T07:01:21.021689Z digest=sha256:a2c33e3da4d394ad928ba6d1753f522085eb6718beeb347f339ca5808e7c6b81

Observation d3862ee3-275f-4aa3-8a1b-9ae35985df42 · outbound

This paper cites Categorizing Variants of Goodhart's Law.

Phantom Guardrails: When Self-Improving Agent Harnesses Fix Failures That Never Happened Categorizing Variants of Goodhart's Law

Reference 13

Resolution
unresolved
no resolver link, observed 2026-08-02T07:01:21.137326Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-02T07:01:21.137326Z digest=sha256:bbaabcb9c2ad4e213dfd86cc0ef3fa2645a579ef6942fc224ac81d34ff7046b8

Observation 949024c6-0e48-453d-b1ed-46b951b70c33 · outbound

This paper cites Code as Agent Harness.

Phantom Guardrails: When Self-Improving Agent Harnesses Fix Failures That Never Happened Code as Agent Harness

Reference 14

Resolution
unresolved
no resolver link, observed 2026-08-02T07:01:21.354832Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-02T07:01:21.354832Z digest=sha256:cfb2f94e05924f889a9ad48a51d5f59c84c247dab5d25e5cc29ab15d1c8eaaa8

Observation 75082944-7ea8-4635-a2f9-65ab8894e5ce · outbound

This paper cites Evolving Agents in the Dark: Retrospective Harness Optimization via Self-Preference.

Phantom Guardrails: When Self-Improving Agent Harnesses Fix Failures That Never Happened Evolving Agents in the Dark: Retrospective Harness Optimization via Self-Preference

Reference 15

Resolution
unresolved
no resolver link, observed 2026-08-02T07:01:21.515063Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-02T07:01:21.515063Z digest=sha256:37905dcf32338e5b58f71f070be7c1812abf5275940b318bd9a8a0e782676f86

Observation b8868949-339a-43c0-835c-d4bee6e7e779 · outbound

This paper cites an unresolved cited work.

Phantom Guardrails: When Self-Improving Agent Harnesses Fix Failures That Never Happened Unresolved cited work

Reference 16

Resolution
unresolved
no resolver link, observed 2026-08-02T07:01:21.678296Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-02T07:01:21.678296Z digest=sha256:4bd6b3fc9e48a9a8eacb03c637d561684b21a2b6877e7222baaf249b7448cdb2

Observation 27ecff90-29e8-47fd-80d3-a35eab35a43d · outbound

This paper cites LLM Agents Already Know When to Call Tools -- Even Without Reasoning.

Phantom Guardrails: When Self-Improving Agent Harnesses Fix Failures That Never Happened LLM Agents Already Know When to Call Tools -- Even Without Reasoning

Reference 17

Resolution
unresolved
no resolver link, observed 2026-08-02T07:01:22.001527Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-02T07:01:22.001527Z digest=sha256:c4e4fb5a4411b4d367f9269ebb03e0c91ccc622f694ea60731cb46273e46aac6

Observation 6c8c60fb-bb4c-4cbc-bdd1-5eb025eeb94b · outbound

This paper cites Beyond Accuracy: Measuring Bias Acknowledgment in Chain-of-Thought Reasoning for Responsible AI Evaluation.

Phantom Guardrails: When Self-Improving Agent Harnesses Fix Failures That Never Happened Beyond Accuracy: Measuring Bias Acknowledgment in Chain-of-Thought Reasoning for Responsible AI Evaluation

Reference 18

Resolution
unresolved
no resolver link, observed 2026-08-02T07:01:22.131040Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-02T07:01:22.131040Z digest=sha256:330d88b1b13132d5856e874036279a5f89a37a6ba7afef8734d8afb90593fa67

Observation af67fb16-db73-489c-a556-c8d266d60bea · outbound

This paper cites VeRO: A Harness for Agents to Optimize Agents.

Phantom Guardrails: When Self-Improving Agent Harnesses Fix Failures That Never Happened VeRO: A Harness for Agents to Optimize Agents

Reference 19

Resolution
unresolved
no resolver link, observed 2026-08-02T07:01:22.313861Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-02T07:01:22.313861Z digest=sha256:352673f557107cecef633f52ce0d92ea99ac71b2dd000bf9e5254d8fa7053892

Observation 4f1244d6-2d8a-4511-8359-fa9953162476 · outbound

This paper cites an unresolved cited work.

Phantom Guardrails: When Self-Improving Agent Harnesses Fix Failures That Never Happened Unresolved cited work

Reference 20

Resolution
unresolved
no resolver link, observed 2026-08-02T07:01:22.439401Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-02T07:01:22.439401Z digest=sha256:e1bfb7c547a3aa4b208741e5e76ed8f8284a08a71cba22be1e6642ce17a12367

Observation f2b6b363-19e1-4878-b8b1-d4691d625625 · outbound

This paper cites Whitson and Adam D.

Phantom Guardrails: When Self-Improving Agent Harnesses Fix Failures That Never Happened Whitson and Adam D

Reference 21

Resolution
malformed identifier
no resolver link, observed 2026-08-02T07:01:22.578939Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-02T07:01:22.578939Z digest=sha256:26ff35673f91cb370ddde0b123882968210faf9dd378b55e59d2db6231c78dbe

Observation f7eb97d0-ab92-4840-aa81-2c947bc2927b · outbound

This paper cites Adapting the Interface, Not the Model: Runtime Harness Adaptation for Deterministic LLM Agents.

Phantom Guardrails: When Self-Improving Agent Harnesses Fix Failures That Never Happened Adapting the Interface, Not the Model: Runtime Harness Adaptation for Deterministic LLM Agents

Reference 22

Resolution
unresolved
no resolver link, observed 2026-08-02T07:01:22.724744Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-02T07:01:22.724744Z digest=sha256:526ea73ac277ba3eee858cd655e0b4af4a2e18143be8455d097f5d83ff90faa3

Observation 2325e54d-7d58-4c2b-9e4d-774778e6dacf · outbound

This paper cites PAFT: Preservation Aware Fine-Tuning for Minimal-Edit Program Repair.

Phantom Guardrails: When Self-Improving Agent Harnesses Fix Failures That Never Happened PAFT: Preservation Aware Fine-Tuning for Minimal-Edit Program Repair

Reference 23

Resolution
unresolved
no resolver link, observed 2026-08-02T07:01:22.844569Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-02T07:01:22.844569Z digest=sha256:59254c7c6bc0e9b1c09ec51b256f05c3928e5fbf2d47646a078faa8a4d12a174

Observation 5056d85c-be0b-4dd2-9a6e-b88e4fdee633 · outbound

This paper cites an unresolved cited work.

Phantom Guardrails: When Self-Improving Agent Harnesses Fix Failures That Never Happened Unresolved cited work

Reference 24

Resolution
unresolved
no resolver link, observed 2026-08-02T07:01:22.927585Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-02T07:01:22.927585Z digest=sha256:d1266c4e573e7557dcaf7740915c2c3823ba4fce6a39014661bd4da49eaafd86

Observation 0e1aaed2-53ac-4a13-8bc1-9161665f35a8 · outbound

This paper cites Reward Auditor: Inference on Reward Modeling Suitability in Real-World Perturbed Scenarios.

Phantom Guardrails: When Self-Improving Agent Harnesses Fix Failures That Never Happened Reward Auditor: Inference on Reward Modeling Suitability in Real-World Perturbed Scenarios

Reference 25

Resolution
unresolved
no resolver link, observed 2026-08-02T07:01:23.147046Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-02T07:01:23.147046Z digest=sha256:cc576017849ba3d6fef82854775b0ad0cead92eca49195ccd53910ea1818bca7

Observation f950cb56-c969-4987-acb1-fb549a5108ee · outbound

This paper cites Defining and Characterizing Reward Hacking.

Phantom Guardrails: When Self-Improving Agent Harnesses Fix Failures That Never Happened Defining and Characterizing Reward Hacking

Reference 2022

Resolution
unresolved
no resolver link, observed 2026-08-02T07:01:21.816420Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-02T07:01:21.816420Z digest=sha256:160cf2c722b2989d59995970c4bb54bc58c98661bf4b82152f00d6d6a8cf9742

Observation 856a6a25-a61c-47ea-be25-8fd08d07a997 · outbound

This paper cites InThe Thirteenth International Conference on Learning Represen- tations.

Phantom Guardrails: When Self-Improving Agent Harnesses Fix Failures That Never Happened InThe Thirteenth International Conference on Learning Represen- tations

Reference 2025

Resolution
unresolved
no resolver link, observed 2026-08-02T07:01:23.034755Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-02T07:01:23.034755Z digest=sha256:c34a2e2f34fa7a879a89ad517479ad8daabbc700bdd2b6bbb19770a3549e19ed

Observation e9f73dff-2960-4a24-8cf7-a864352ab80a · outbound

This paper cites QiMeng-PRepair: Precise Code Repair via Edit-Aware Reward Optimization.

Phantom Guardrails: When Self-Improving Agent Harnesses Fix Failures That Never Happened QiMeng-PRepair: Precise Code Repair via Edit-Aware Reward Optimization

Reference 2026

Resolution
unresolved
no resolver link, observed 2026-08-02T07:01:19.894262Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

source=pdf_text observed=2026-08-02T07:01:19.894262Z digest=sha256:9aee76ed2cf028affbfef2dbfd11ecb1d505244138ac988a86463ad1d40ca0f3

Pith citing papers

No inbound Pith citation observations are available.