Security context
What an agent needs to avoid regressing past fixes and find the next vuln in this repo.
rtk-ai/rtk
develop @ ee9e2f8
27
Fixes
4
CVEs
HIGH
Peak severity
47.6%
Coverage
Highlights
Auth Bypass: 16 prior fixes. Scrutinize any change in this area.
src/hooks/permissions.rs: most-fixed (5 issues). Treat as high-risk during review.
19 high-severity fixes in this history; regressions here are high-impact.
Recurring patterns
The bug types that recur here, drawn from past fixes, not open vulnerabilities.
Auth Bypass: The project's central command permission engine has suffered multiple authorization bypasses. Complex compound shell commands, process substitutions, and file redirects have historically allowed unauthorized actions to slip past validation rules due to incomplete command parsing.
Remote Code Execution: Trusting project-local configurations allows attackers to place malicious .rtk/filters.toml files in a working directory to modify command outputs and hide unauthorized code execution. Preventing this requires enforcing a strict signature verification layer.
Command Injection: The OpenClaw plugin used unescaped template strings with shell-executing commands, permitting shell command injection when argument array execution was not used.