Command Injection
A shell is an interpreter: when user data becomes part of a command string, punctuation can become behavior.
Frame the problem
Security starts with a concrete asset, attacker capability and trust crossing.
Why the system fails
The program invokes a shell with a concatenated string instead of a direct API or fixed executable plus argument array.
The important question is not “what is Command Injection?” but “which assumption let untrusted data or an over-scoped identity cross application value → os shell?” Trace the decision at the boundary, then constrain what can happen after the first control fails.
Design the control in layers
Start with the control closest to the interpretation or privilege boundary: Avoid the shell; call a safe library or executable with an argument array Then add a control that reduces blast radius and telemetry that proves the decision was enforced.
The resulting design is not labelled secure. Record the identified controls, the known failure paths, the remaining exposure, and the evidence you would need during an incident.
| Prevent | Detect | Recover |
|---|---|---|
| Avoid the shell; call a safe library or executable with an argument array · Allowlist valid values · Run as a restricted user in a sandbox | Unexpected child processes and command lines · Egress from a process that normally has none | Contain the affected identity or component, scope impact from audit evidence, and preserve a regression test. |
Key points
- Asset: Host files, credentials, process integrity and downstream network access.
- Boundary: Application value → OS shell
- Primary control: Avoid the shell; call a safe library or executable with an argument array
- Detection signal: Unexpected child processes and command lines
- Always ask what limits damage when the primary control fails.
Boundary control exercise
This lesson uses the shared boundary-control exercise.
Follow the attack
Safe conceptual simulation: capability → missing control → crossed boundary → asset impact.
- 1Attacker starts with: A caller who controls a filename, option or command fragment.
- 2The program invokes a shell with a concatenated string instead of a direct API or fixed executable plus argument array.
- 3The weak or missing boundary control is crossed: Application value → OS shell
- 4Impact: Code execution with the service identity and everything that identity can reach.
- Code execution with the service identity and everything that identity can reach.
Defend, detect, recover
One prevention is a single point of security failure. Layer it and make failure observable.
- • Avoid the shell; call a safe library or executable with an argument array
- • Allowlist valid values
- • Run as a restricted user in a sandbox
- • Unexpected child processes and command lines
- • Egress from a process that normally has none
- • Contain the affected identity or component.
- • Scope access from audit evidence.
- • Fix the boundary and add a regression test.
- • Misconfiguration and new access paths can bypass the intended control.
- • A privileged insider or compromised control plane may still reach the asset.