@inproceedings{yan-etal-2026-riskguard,
title = "{R}isk{G}uard: Hypergraph-Based Pre-Execution Risk Control for Long-Horizon Tool Use",
author = "Yan, Yandong and
Cai, Zesen and
Shi, Daming and
He, Yifei and
Dai, Ruiting and
Huang, Yu",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {EMNLP} 2026",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-emnlp.2066/",
pages = "43979--43997",
ISBN = "979-8-89176-462-0",
abstract = "In long-horizon tool use, early-step defects passing per-call checks often propagate silently across dependency chains to cause irreversible late-stage failures{---}a propagation class accounting for 73.0{\%} of diagnosed Vanilla FC failures. Because these defects are locally valid (stepwise legality) yet manifest differently across tool boundaries (semantic heterogeneity), they easily evade standard per-call guardrails and post-hoc reflection, necessitating proactive pre-execution risk anticipation. To this end, we formalize the pre-execution risk propagation problem and propose RiskGuard, a training-free, dev-calibrated runtime controller. To capture the multi-source, conjunctive dependencies inherent in tool composition, RiskGuard models execution traces via a typed dependency hypergraph. Meanwhile, a Referential/Constraint/Operational (R/C/O) kernel tracks risk transitions across semantic boundaries. Replacing rigid binary blocking, a five-mode intervention rule selects the lowest-commitment feasible response to mitigate risks smoothly. Evaluated on ToolSandbox-hard, pooled {\ensuremath{\tau}}{\texttwosuperior}-Bench, and ToolHop, RiskGuard achieves absolute gains of +0.133 Sim (0.281 {\textrightarrow} 0.414), +9.2 pp pass@1, and +4.85 pp accuracy over the baseline, with depth diagnostics showing stronger gains on harder dependency slices."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="yan-etal-2026-riskguard">
<titleInfo>
<title>RiskGuard: Hypergraph-Based Pre-Execution Risk Control for Long-Horizon Tool Use</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yandong</namePart>
<namePart type="family">Yan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zesen</namePart>
<namePart type="family">Cai</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Daming</namePart>
<namePart type="family">Shi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yifei</namePart>
<namePart type="family">He</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ruiting</namePart>
<namePart type="family">Dai</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yu</namePart>
<namePart type="family">Huang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: EMNLP 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hung-Yi</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunipa</namePart>
<namePart type="family">Dev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jessy</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Desmond</namePart>
<namePart type="family">Elliott</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-462-0</identifier>
</relatedItem>
<abstract>In long-horizon tool use, early-step defects passing per-call checks often propagate silently across dependency chains to cause irreversible late-stage failures—a propagation class accounting for 73.0% of diagnosed Vanilla FC failures. Because these defects are locally valid (stepwise legality) yet manifest differently across tool boundaries (semantic heterogeneity), they easily evade standard per-call guardrails and post-hoc reflection, necessitating proactive pre-execution risk anticipation. To this end, we formalize the pre-execution risk propagation problem and propose RiskGuard, a training-free, dev-calibrated runtime controller. To capture the multi-source, conjunctive dependencies inherent in tool composition, RiskGuard models execution traces via a typed dependency hypergraph. Meanwhile, a Referential/Constraint/Operational (R/C/O) kernel tracks risk transitions across semantic boundaries. Replacing rigid binary blocking, a five-mode intervention rule selects the lowest-commitment feasible response to mitigate risks smoothly. Evaluated on ToolSandbox-hard, pooled \ensuremathτ²-Bench, and ToolHop, RiskGuard achieves absolute gains of +0.133 Sim (0.281 → 0.414), +9.2 pp pass@1, and +4.85 pp accuracy over the baseline, with depth diagnostics showing stronger gains on harder dependency slices.</abstract>
<identifier type="citekey">yan-etal-2026-riskguard</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.2066/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>43979</start>
<end>43997</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T RiskGuard: Hypergraph-Based Pre-Execution Risk Control for Long-Horizon Tool Use
%A Yan, Yandong
%A Cai, Zesen
%A Shi, Daming
%A He, Yifei
%A Dai, Ruiting
%A Huang, Yu
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Findings of the Association for Computational Linguistics: EMNLP 2026
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-462-0
%F yan-etal-2026-riskguard
%X In long-horizon tool use, early-step defects passing per-call checks often propagate silently across dependency chains to cause irreversible late-stage failures—a propagation class accounting for 73.0% of diagnosed Vanilla FC failures. Because these defects are locally valid (stepwise legality) yet manifest differently across tool boundaries (semantic heterogeneity), they easily evade standard per-call guardrails and post-hoc reflection, necessitating proactive pre-execution risk anticipation. To this end, we formalize the pre-execution risk propagation problem and propose RiskGuard, a training-free, dev-calibrated runtime controller. To capture the multi-source, conjunctive dependencies inherent in tool composition, RiskGuard models execution traces via a typed dependency hypergraph. Meanwhile, a Referential/Constraint/Operational (R/C/O) kernel tracks risk transitions across semantic boundaries. Replacing rigid binary blocking, a five-mode intervention rule selects the lowest-commitment feasible response to mitigate risks smoothly. Evaluated on ToolSandbox-hard, pooled \ensuremathτ²-Bench, and ToolHop, RiskGuard achieves absolute gains of +0.133 Sim (0.281 → 0.414), +9.2 pp pass@1, and +4.85 pp accuracy over the baseline, with depth diagnostics showing stronger gains on harder dependency slices.
%U https://aclanthology.org/2026.findings-emnlp.2066/
%P 43979-43997
Markdown (Informal)
[RiskGuard: Hypergraph-Based Pre-Execution Risk Control for Long-Horizon Tool Use](https://aclanthology.org/2026.findings-emnlp.2066/) (Yan et al., Findings 2026)
ACL