Back to Directory
59

SWE-agent

AI Coding Agent

by Princeton NLP

Autonomous agent that solves real GitHub issues. Uses LLMs to navigate codebases, understand bugs, and submit pull requests. Research tool from Princeton.

Updated 2026-03-09
View Source

Trust Score Breakdown

Security(30%)
70/100
Source code available for audit1 high-risk permission(s): Execute commands2 medium-risk permission(s)Permissions explicitly declaredVersion 0.7.0 — pre-stableAccountable vendor with web presence
Reliability(30%)
57/100
Good documentation providedVersion 0.7.0 — pre-1.0 developmentSource repository linked — enables issue trackingActive project website — ongoing support signalUpdated within the last week
Transparency(20%)
83/100
Open source — full code transparencyOfficial website with project informationDetailed description providedPermissions declaredRelease version tracked
Community(10%)
0/100
No GitHub repository linkedNo npm download data

Permissions Required

File access

Network access

Execute commands

Reviews (0)

Sign in to leave a review.

No reviews yet. Be the first to review this agent.