autonomous-coding-agents
Warn
Audited by Socket on Oct 13, 2026
1 alert found:
AnomalyAnomalyreferences/claude-code-cli-legacy.md
LOWAnomalyLOW
references/claude-code-cli-legacy.md
SUSPICIOUS: the skill’s core purpose matches Claude Code orchestration, and the main CLI install path is official, but it disproportionately instructs the agent to bypass safety prompts and automate consent for `--dangerously-skip-permissions`. The result is a medium-high security risk from autonomy and external tool execution rather than confirmed malware.
Confidence: 90%Severity: 62%
Audit Metadata