Skip to content

Commit abbac96

Browse files
committed
Event: add talk by Haoyu Wang@19Aug
1 parent 2909e47 commit abbac96

File tree

1 file changed

+65
-0
lines changed

1 file changed

+65
-0
lines changed

content/event/250819.md

Lines changed: 65 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,65 @@
1+
---
2+
title: "AgentSpec: Customizable Runtime Enforcement for Safe and Reliable LLM Agents"
3+
4+
event: Weekly Talk
5+
event_url:
6+
7+
location: COM3-B1-15 - Meeting Rm 92
8+
address:
9+
street:
10+
city:
11+
region:
12+
postcode:
13+
country: Singapore
14+
15+
summary:
16+
abstract: "Agents built on LLMs are increasingly deployed across diverse domains, automating complex decision-making and task execution. However, their autonomy introduces safety risks, including security vulnerabilities, legal violations, and unintended harmful actions. Existing mitigation methods, such as model-based safeguards and early enforcement strategies, fall short in robustness, interpretability, and adaptability. To address these challenges, we propose AgentSpec, a lightweight domain-specific language for specifying and enforcing runtime constraints on LLM agents. With AgentSpec, users define structured rules that incorporate triggers, predicates, and enforcement mechanisms, ensuring agents operate within predefined safety boundaries. We implement AgentSpec across multiple domains, including code execution, embodied agents, and autonomous driving, demonstrating its adaptability and effectiveness. Our evaluation shows that AgentSpec successfully prevents unsafe executions in over 90% of code agent cases, eliminates all hazardous actions in embodied agent tasks, and enforces 100% compliance by autonomous vehicles (AVs). Despite its strong safety guarantees, AgentSpec remains computationally lightweight, with overheads in milliseconds. By combining interpretability, modularity, and efficiency, AgentSpec provides a practical and scalable solution for enforcing LLM agent safety across diverse applications. We also automate the generation of rules using LLMs and assess their effectiveness. Our evaluation shows that the rules generated by OpenAI o1 achieve a precision of 95.56% and recall of 70.96% for embodied agents, successfully identify 87.26% of the risky code, and prevent AVs from breaking laws in 5 out of 8 scenarios."
17+
18+
# Talk start and end times.
19+
# End time can optionally be hidden by prefixing the line with `#`.
20+
date: "2025-08-19T16:00:00Z"
21+
date_end: "2025-08-19T17:00:00Z"
22+
all_day: false
23+
24+
# Schedule page publish date (NOT talk date).
25+
publishDate: "2017-01-01T00:00:00Z"
26+
27+
authors: [Haoyu Wang]
28+
tags: [Weekly Talk]
29+
30+
# Is this a featured talk? (true/false)
31+
featured: false
32+
33+
image:
34+
caption: 'Image credit: [**Unsplash**](https://unsplash.com/photos/bzdhc5b3Bxs)'
35+
focal_point: Right
36+
37+
url_code: ""
38+
url_pdf: ""
39+
url_slides: ""
40+
url_video: ""
41+
42+
# Markdown Slides (optional).
43+
# Associate this talk with Markdown slides.
44+
# Simply enter your slide deck's filename without extension.
45+
# E.g. `slides = "example-slides"` references `content/slides/example-slides.md`.
46+
# Otherwise, set `slides = ""`.
47+
slides:
48+
49+
# Projects (optional).
50+
# Associate this post with one or more of your projects.
51+
# Simply enter your project's folder or file name without extension.
52+
# E.g. `projects = ["internal-project"]` references `content/project/deep-learning/index.md`.
53+
# Otherwise, set `projects = []`.
54+
projects:
55+
56+
# Slides can be added in a few ways:
57+
#
58+
# - **Create** slides using Wowchemy's [*Slides*](https://wowchemy.com/docs/managing-content/#create-slides) feature and link using `slides` parameter in the front matter of the talk file
59+
# - **Upload** an existing slide deck to `static/` and link using `url_slides` parameter in the front matter of the talk file
60+
# - **Embed** your slides (e.g. Google Slides) or presentation video on this page using [shortcodes](https://wowchemy.com/docs/writing-markdown-latex/).
61+
#
62+
# Further event details, including page elements such as image galleries, can be added to the body of this page.
63+
64+
65+
---

0 commit comments

Comments
 (0)