autonomous-agent-harness
Transforms Claude Code into a fully autonomous agent system with persistent memory and scheduled tasks for continuous operation.
Install this skill
or
80/100
Security score
The autonomous-agent-harness skill was audited on May 15, 2026 and we found 8 security issues across 3 threat categories. Review the findings below before installing.
Categories Tested
Security Issues
medium line 123
Curl to non-GitHub URL
SourceSKILL.md
| 123 | curl -X POST "https://api.anthropic.com/dispatch" \ |
low line 127
Webhook reference - potential data exfiltration
SourceSKILL.md
| 127 | # Trigger from webhook |
low line 128
Webhook reference - potential data exfiltration
SourceSKILL.md
| 128 | # GitHub webhook → dispatch → Claude agent → fix → PR |
medium line 59
Access to hidden dotfiles in home directory
SourceSKILL.md
| 59 | **内置内存**（`~/.claude/projects/*/memory/`）： |
low line 78
Access to hidden dotfiles in home directory
SourceSKILL.md
| 78 | 写入 ~/.claude/projects/*/memory/ 以实现跨会话回忆 |
low line 159
Access to hidden dotfiles in home directory
SourceSKILL.md
| 159 | 将任务队列写入 ~/.claude/projects/*/memory/task-queue.md |
medium line 193
Access to hidden dotfiles in home directory
SourceSKILL.md
| 193 | 确保这些在 `~/.claude.json` 中： |
low line 123
External URL reference
SourceSKILL.md
| 123 | curl -X POST "https://api.anthropic.com/dispatch" \ |
Scanned on May 15, 2026
View Security DashboardGitHub Stars 189.4K
Rate this skill
Categorydevelopment
UpdatedMay 25, 2026
affaan-m/everything-claude-code