strands-evals

Warn

Audited by Socket on Sep 24, 2026

2 alerts found:

Anomalyx2
AnomalyLOW
.github/workflows/strands-command.yml

No direct malicious behavior is evident in this workflow fragment. The primary concern is supply-chain exposure from third-party actions pinned to mutable @main branches, particularly actions running with write permissions or access to OIDC and AWS secrets. Dispatch authorization relies on GitHub's workflow-dispatch access controls because the explicit authorization check is skipped.

Confidence: 96%Severity: 57%
AnomalyLOW
.github/workflows/issue-labeler.yml

No direct malicious behavior is visible in this workflow, but its use of pull_request_target with a mutable third-party action, write permissions, and OIDC capability creates meaningful supply-chain and privilege-escalation risk. Pin actions to verified commit SHAs, minimize permissions, and confirm the AWS role trust policy and action implementation.

Confidence: 95%Severity: 68%
Audit Metadata
Analyzed At
Sep 24, 2026, 12:03 AM
Package URL
pkg:socket/skills-sh/strands-agents%2Fevals%2Fstrands-evals%2F@9ea55c96d461d70b0d742f547a95389e4c7fe18d