The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill downloads standard Python libraries (click, python-gitlab, requests) from the official Python Package Index (PyPI). It also fetches workflow artifacts from GitHub's official infrastructure using the gh CLI tool.
[COMMAND_EXECUTION]: The skill executes local repository scripts (download_golden_values.py and compare_golden_values_kl.py) and standard shell commands for environment setup and file cleanup.
[CREDENTIALS_UNSAFE]: The skill manages authentication by retrieving a temporary token via the gh auth token command. It includes specific instructions to avoid committing or storing this token in long-lived environments, following security best practices for one-shot command execution.
[PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it processes external data from GitHub workflow artifacts to generate a PR summary.
Ingestion points: External data enters the system through JSON artifacts downloaded into tests/functional_tests/test_cases/.
Boundary markers: The skill uses a predefined Markdown template for the summary blurb, which helps structure the output but does not include explicit boundary markers for the data segments.
Capability inventory: The skill has the capability to overwrite local repository files (golden values) and perform network operations via the python-gitlab and requests libraries.
Sanitization: The provided instructions do not detail specific sanitization logic for the artifact content, relying on the internal Python scripts for processing.

update-golden-values