tdd-repair
Warn
Audited by Socket on Jun 25, 2026
1 alert found:
AnomalyAnomalySKILL.md
LOWAnomalyLOW
SKILL.md
SUSPICIOUS: the skill is internally coherent for automated test-driven repair and uses the official Claude CLI, so it is not strong evidence of malware. However, it grants a nested agent Bash+Edit capability over untrusted repo content and executes caller-supplied test commands, creating meaningful code-execution and prompt-injection risk that is disproportionate for untrusted projects.
Confidence: 84%Severity: 58%
Audit Metadata