You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
command-injection: decide on a distinct argument-injection rule for option cases #220
#208 (merged in #217) reports option injection, an attacker-controlled element of an argument list run without a shell, under the rule id command-injection and the message title Command injection, with the case in the evidence (…the input is a whole element, so it may be read as an option). The issue left the naming to a later decision.
Options:
A distinct rule argument-injection for the option cases (4, 5b, 5c, 7 of command-injection: severity of option injection in argument lists run without a shell #208), command-injection keeping the shell, program, operand-command and interpreter cases. Clearer for users and for suppressions and baselines, but a new rule id: existing suppressions and baselines of those findings would no longer match, and the SARIF rule catalogue grows.
Keep command-injection with a distinct message title (Option injection) and a metadata field (injection: option | command). No change to rule ids, suppressions or baselines; filtering needs the metadata.
To settle
Which option, and for the first: the rule's default severity, its SARIF description, and whether a release note announces the split for existing suppressions and baselines.
Activity
Merge pull request #232 from CoreTrace/feat/command-injection-nature