The Agent Skills Directory

[NO_CODE]: The scripts/README.md file lists three Python scripts (test_selector.py, test_executor.py, and report_generator.py) as bundled resources, but the actual source code for these files is missing from the skill package, preventing a verification of their safety.
[DYNAMIC_EXECUTION]: The skill instructions require the agent to generate and write pipeline configuration code into CI/CD files (e.g., .github/workflows/test.yml), which involves creating executable logic based on templates.
[INDIRECT_PROMPT_INJECTION]: The skill processes and aggregates data from external, potentially untrusted test result files (JUnit XML, JSON), which is a known attack surface for indirect prompt injection.
Ingestion points: Step 6 in SKILL.md and the report_generator.py script are designed to read and merge external XML and JSON test results.
Boundary markers: There are no specific instructions or delimiters defined to ensure the agent ignores or sanitizes instructions that might be embedded within test names or error messages.
Capability inventory: The skill possesses Write and Edit permissions for project files and restricted Bash access for test execution.
Sanitization: The instructions do not specify any validation or sanitization steps for the data extracted from the test results.

orchestrating-test-execution