ci/sast
Code is scanned by a SAST tool
A capability: each stack checks it with its own tools. Implemented for any, php, ruby.
Combine: all. Every implementation that runs must pass: each stack in the repository meets the requirement with its own tools.
| Severity | info |
| Tiers | 1 |
| Stacks | any |
| Scope | root |
| Kind | capability |
| Source | library v0.6.0 |
| Selected by | fleetlint-recommended@v0.6.0 |
Requirement
A static application security test runs in CI or the task runner: a stack-neutral scanner (CodeQL, Semgrep or Opengrep, GitLab SAST, SonarQube) or the stack’s own taint analysis (Brakeman for Rails, Psalm’s taint analysis for PHP). Linters with security rules count under the lint rules, not here. CodeQL default setup, which leaves no file in the repository, and other scanners are declared with accept:.
Why
SAST follows data from input to dangerous sinks across files, which per-file linters do not; on GitHub CodeQL is free for public repositories.
Fix
Add a CodeQL workflow (GitHub advanced setup), include the GitLab SAST component, or add a Semgrep step to CI; with CodeQL default setup, declare it with accept:.
Agent instruction: Add a CodeQL workflow (github/codeql-action init/analyze, pinned) for the repository’s languages, or on GitLab include components/sast; run it on pull requests and weekly.
Sources
- https://github.com/ossf/scorecard/blob/main/docs/checks.md (standard, SAST)
- https://csrc.nist.gov/pubs/sp/800/218/final (standard, PW.7.2: Perform the code review and/or code analysis based on the organization’s secure coding standards)
- https://fleetlint.org/baseline/#security
Generated by fleetlint docs from catalog v0.6.0.