The OpenAI Agents SDK is a lightweight yet powerful framework for building multi-agent workflows.
88%
Total Score
healthy
Active releases, broad organization backing, and fresh repository activity outweigh concentrated commits and workflow audit concerns.
The 18 publishing accounts are predominantly on openai.com, consistent with the organization-backed repository. One gmail.com publishing account, aibrahim-openai, is an account-hygiene caution rather than evidence of limited maintenance capacity.
The repository name does not match the package name and its README does not mention this package. This is a caution that the linkage is not explicit, although the shared organization and related package structure provide some context.
Version 0.20.0 is not a stable-major release, so API compatibility may still change, but it is not a prerelease and recent releases contain no prerelease versions.
All 8 workflows were analyzed successfully; actions are fully pinned, permissions are read-only or job-scoped, and no untrusted checkout or script injection was found. A high-confidence template-injection finding and trusted-publishing audit result warrant workflow caution, though the latter is informational and no broad write permissions are present.
| Title | Versions | Severity |
|---|---|---|
AIKIDO-2026-63881 Pre-CVE Found by Aikido Intel before public disclosure or CVE publication. @openai/agents-openai is vulnerable to Improper Authorization in versions 0.0.7 - 0.10.0. | 0.0.7 - 0.10.0 | Medium |
| Dependency | Last Release | Score |
|---|---|---|
openai Version ^7.2.0 | — | — |
@openai/agents-core Version 0.20.0 | — | — |
Connect your repositories to instantly see whether vulnerable or malicious packages exist in your codebase.
Free. No credit card required.

I consent to receiving marketing communications based on Aikido’s Privacy Policy.
SOC 2Compliant
ISO 27001Compliant
ISO 42001Compliant© All Intel data is openly available and commercially licensed.