Feature: HarmActEval benchmark for agent tool calls#206
Open
prane-eth wants to merge 2 commits intoconfident-ai:mainfrom
Open
Feature: HarmActEval benchmark for agent tool calls#206prane-eth wants to merge 2 commits intoconfident-ai:mainfrom
prane-eth wants to merge 2 commits intoconfident-ai:mainfrom
Conversation
|
@prane-eth is attempting to deploy a commit to the Confident AI Team on Vercel. A member of the Team first needs to authorize it. |
prane-eth
changed the title
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Solves #202
Background
AI agents have a growing adoption across the industry, including critical applications. AI agents that have access to tools (including MCP servers) can currently call tools directly with no centralized validation layer that inspects these calls before execution, allowing harmful or disallowed tool calls to be executed without oversight.
Agent-Action-Guard introduced HarmActEval to evaluate harmful actions by agents based on harmful instructions and availability of harmful tools. The HarmActEval experiments proved GPT-5.3 has a safety score of 17.33%, which shows a very high vulnerability, proving the requirement for the evaluation. Agent-Action-Guard received 605 downloads on PyPI and 247 clones on GitHub in the first week.
Changes
I integrated HarmActEval to allow evaluation of agent safety at action-level, which was not covered by any other research or benchmarks. This allows to identify specific harmful actions taken by agents. HarmActEval is a dataset-based evaluation benchmark.
I added the dataset, benchmark, test cases, and documentation. All the test cases got passed.