# Beyond Aggregate Risk: Role-Stratified Conformal Risk Control for LLM Tool Calls

arXiv: https://arxiv.org/abs/2607.24343 · HTML: https://mlsanity.com/paper/2607.24343/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Agent threats, guardrails and governance, one of 16 directions in LLM agents and tool use.
- Builds on 21 of its 36 references that are in the map.
- First posted on arXiv 27 Jul 2026.

## Placed in

- LLM agents and tool use / [Agent threats, guardrails and governance](https://mlsanity.com/agents/agent-threats-guardrails-and-governance/)

## Builds on (most cited first, 21 of 21)

- [The Llama 3 Herd of Models](https://mlsanity.com/paper/2407.21783/) (2024)
- [Universal and Transferable Adversarial Attacks on Aligned Language Models](https://mlsanity.com/paper/2307.15043/) (2023)
- [Distribution-Free Predictive Inference For Regression](https://mlsanity.com/paper/1604.04173/) (2016)
- [The Instruction Hierarchy: Training LLMs to Prioritize Privileged Instructions](https://mlsanity.com/paper/2404.13208/) (2024)
- [StruQ: Defending Against Prompt Injection with Structured Queries](https://mlsanity.com/paper/2402.06363/) (2024)
- [Defeating Prompt Injections by Design](https://mlsanity.com/paper/2503.18813/) (2025)
- [Distribution-Free, Risk-Controlling Prediction Sets](https://mlsanity.com/paper/2101.02703/) (2021)
- [Learn then Test: Calibrating Predictive Algorithms to Achieve Risk Control](https://mlsanity.com/paper/2110.01052/) (2021)
- [Defending Against Indirect Prompt Injection Attacks With Spotlighting](https://mlsanity.com/paper/2403.14720/) (2024)
- [SecAlign: Defending Against Prompt Injection with Preference Optimization](https://mlsanity.com/paper/2410.05451/) (2024)
- [Class-Conditional Conformal Prediction with Many Classes](https://mlsanity.com/paper/2306.09335/) (2023)
- [Jatmo: Prompt Injection Defense by Task-Specific Finetuning](https://mlsanity.com/paper/2312.17673/) (2023)
- [Distribution-free uncertainty quantification for classification under label shift](https://mlsanity.com/paper/2103.03323/) (2021)
- [Kandinsky Conformal Prediction: Beyond Class- and Covariate-Conditional Coverage](https://mlsanity.com/paper/2502.17264/) (2025)
- [The Granularity Mismatch in Agent Security: Argument-Level Provenance Solves Enforcement and Isolates the LLM Reasoning Bottleneck](https://mlsanity.com/paper/2605.11039/) (2026)
- [Taxonomy, Evaluation and Exploitation of IPI-Centric LLM Agent Defense Frameworks](https://mlsanity.com/paper/2511.15203/) (2025)
- [Conformal Selective Acting: Anytime-Valid Risk Control for RLVR-Trained LLMs](https://mlsanity.com/paper/2605.20270/) (2026)
- [CORA: Conformal Risk-Controlled Agents for Safeguarded Mobile GUI Automation](https://mlsanity.com/paper/2604.09155/) (2026)
- [AgentVisor: Defending LLM Agents Against Prompt Injection via Semantic Virtualization](https://mlsanity.com/paper/2604.24118/) (2026)
- [ToolChain-CRC: Conformal Risk Control for Agentic AI Under Retrieval and Tool-Use Drift](https://mlsanity.com/paper/2606.18467/) (2026)
- [Taming Variability: Randomized and Bootstrapped Conformal Risk Control for LLMs](https://mlsanity.com/paper/2509.23007/) (2025)

## Built on by (most cited first, 1 of 1)

- [Finding Blind Spots in AppWorld and WorkArena Task Verifiers](https://mlsanity.com/paper/2610.09142/) (2026)

## Abstract (arXiv)

Language-model agents act through structured tool calls whose arguments carry very different risks: untrusted content may legitimately shape an email body but should never set a recipient, account, command, or credential. Existing conformal risk control methods certify a tool call as a whole, so a failure in one rare high-risk field can be averaged away by the many benign arguments around it, leaving the argument that causes harm uncertified. We introduce role-stratified per-field conformal risk control, a calibration layer that wraps any per-field detector and assigns a separate threshold and risk budget to each semantic argument role. We show that aggregate certification pays a price of coarseness, tightening a rare role's effective budget in proportion to how often that role appears, whereas role-stratified calibration certifies each sufficiently sampled role directly with a finite-sample guarantee and pools the rarest roles. Across AgentDojo and InjecAgent with six language models, our method achieves the most consistent role-specific budget compliance among the methods we evaluate under model and attack transfer, detector noise, gradual drift, unseen tool suites, and adaptive attacks, providing formal per-role guarantees under exchangeability or after recalibration. These results suggest that structured tool calls should be certified at the semantic-role level, not the whole action.