Sanitizer boundary: Input reaches command grammar where shell or process syntax can change what executes.
The pack does not provide a callable signature for this record; the watched access path is shown directly.
| Arg | Access path | Purpose | Watched |
|---|---|---|---|
| Argument[0] | Argument[0] | The access path Atropos marks for this model. | ▲ sanitizer |
CWE-78 ↗
OS Command Injection
External input alters the syntax or behavior of an operating-system command.
Attacker's-eye view · command-injection
An attacker adds separators, substitutions, or flags that turn a value into a second command or alter the first.
// modeled boundary: Argument[0] -> ReturnValue
quote(attacker_influenced_value);
1Use a fixed executable and argument array.
2Keep shell parsing disabled where possible.
3Validate executable and argument domains.
These checks require the surrounding codebase. The model names the boundary; it does not decide reachability or prove that a guard dominates every path.
Is this a bug in your code?Atropos stops here — by design
Atropos identifies Argument[0] -> ReturnValue as a command-injection sanitizer. It cannot see whether untrusted data reaches this call in your repository.
Lachesis is the codebase-level step: it traces reachability and guards for this symbol.
Check this symbol in Lachesis →Model facts · verbatim from the pack
| Role | Kind | Access path | Model ID | Confidence |
|---|---|---|---|---|
| sanitizer | command-injection | Argument[0] -> ReturnValue | python.shlex.quote.a0-ret | high · corrob. 3 |