autoresearch-agent
Warn
Audited by Socket on May 5, 2026
1 alert found:
AnomalyAnomalySKILL.md
LOWAnomalyLOW
SKILL.md
SUSPICIOUS: the skill's core behavior is coherent with its stated purpose, but it grants an agent a high-autonomy edit/execute/git loop that can run indefinitely and use arbitrary evaluation commands. Install provenance is mostly consistent for the GitHub repo, with moderate extra risk from the unverified registry-based install path; no clear credential harvesting or malicious exfiltration is shown.
Confidence: 84%Severity: 58%
Audit Metadata