The Agent Skills Directory

[SAFE]: The skill performs legitimate file system operations consistent with its documented purpose. It reads repository metadata (such as package.json, requirements.txt, and go.mod) to identify the project stack and writes generated YAML configuration files to local paths. All operations are local and do not involve network exfiltration or credential access.
[PROMPT_INJECTION]: The skill is subject to a surface for indirect prompt injection because it processes untrusted data from repository configuration files. However, the risk is minimal as the skill does not execute the processed data and only uses it to populate static templates for user review.
Ingestion points: The scripts scripts/stack_detector.py and scripts/pipeline_generator.py read the scripts section from package.json and other manifest files.
Boundary markers: None present.
Capability inventory: The skill utilizes Path.write_text in scripts/pipeline_generator.py to save the generated pipelines.
Sanitization: No explicit sanitization or escaping of the values extracted from repository files is performed before they are interpolated into the CI/CD templates.

ci-cd-pipeline-builder