setup-crap-check-github-actions
Pass
Audited by Gen Agent Trust Hub on May 5, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The Go language preset generates a workflow that installs the
gcov2lcovtool directly from a third-party GitHub repository (github.com/jandelgado/gcov2lcov@latest). - [COMMAND_EXECUTION]: The skill generates a GitHub Actions YAML file (
.github/workflows/crap-check.yml) that executes shell commands for testing, coverage reporting, and CRAP score calculation. These commands (e.g.,mvn test,npm ci,pytest) are executed in the CI environment. - [PROMPT_INJECTION]: The skill exhibits a surface for Indirect Prompt Injection (Category 8):
- Ingestion points: The skill reads project configuration files (like
package.json,pom.xml, or.coveragerc) to automatically detect test frameworks and command-line arguments inSKILL.md(Step 1). - Boundary markers: Absent. The detected strings are passed as arguments to the
generate_workflow.pyscript and interpolated directly into the workflow YAML. - Capability inventory: The skill has the capability to write files to the
.github/workflows/directory and include arbitrary shell commands in the generated workflow. - Sanitization: None. The skill relies on a manual user confirmation step ("Present findings to user for confirmation before proceeding") to mitigate the risk of injecting malicious commands into the CI pipeline.
Audit Metadata