[codex] Check UDF skill markdown license headers

[WilliamK112]

Fixes #15112.

Description

This adds skills/udf-*/*.md to the license-header workflow include patterns so changes to UDF skill markdown files are checked for a current-year NVIDIA copyright/license header. The pattern is scoped to UDF skill content and avoids broad *.md coverage, which would include repository docs that do not use skill front matter.

Validation:

• git diff --check
• Parsed .github/workflows/license-header-check.yml with Ruby YAML
• Locally emulated the license-header action glob for all skills/udf-* markdown files
• Confirmed skills/README.md and skills/docs/dev/VERSIONS.md are not matched by the new pattern

Checklists

Documentation

• Updated for new or modified user-facing features or behaviors
• No user-facing change

Testing

• Added or modified tests to cover new code paths
• Covered by existing tests
  (Please provide the names of the existing tests in the PR description.)
• Not required

Performance

• Tests ran and results are added in the PR description
• Issue filed with a link in the PR description
• Not required

[greptile-apps]

Greptile Summary

This PR adds skills/udf-*/*.md to the included_file_patterns list in the license-header workflow so that UDF skill markdown files are checked for a current-year NVIDIA copyright header on every PR, and removes a trailing blank line at the end of the file.

• The glob skills/udf-*/*.md matches exactly one directory level, leaving markdown files nested in subdirectories (e.g. skills/udf-convert-to-cuda/references/*.md) outside the check — those files currently have headers, but the gap exists for future additions.
• The seven SKILL.md files encode copyright in YAML front matter (metadata.spdx-file-copyright-text) rather than in an HTML comment block; whether the NVIDIA/spark-rapids-common/license-header-check action recognises this format needs confirmation to avoid CI failures on future PRs that touch those files.

Confidence Score: 4/5

The workflow change is minimal and targeted; the main risk is that SKILL.md files use a non-comment copyright format that may not satisfy the action, which could silently skip enforcement or cause CI noise on future unrelated PRs.

The glob addition is correct for the top-level udf-* markdown files and the trailing-newline cleanup is harmless. The open question is whether the license-header action handles the YAML front-matter copyright format used by SKILL.md files — if it does not, every subsequent PR touching a SKILL.md will fail the check unexpectedly.

.github/workflows/license-header-check.yml — confirm the action accepts YAML front-matter copyright and consider whether the pattern should cover subdirectory markdown files.

Important Files Changed

Filename	Overview
.github/workflows/license-header-check.yml	Adds skills/udf-*/*.md to the license-header-check include list and removes a trailing blank line; pattern covers only immediate children of udf-* directories, leaving subdirectory markdown files (references/*.md) unmatched.

Flowchart

%%{init: {'theme': 'neutral'}}%%
flowchart TD
    A[PR opened / synchronized] --> B[license-header-check workflow]
    B --> C{File matches included_file_patterns?}
    C -->|*.yml, *.scala, *.py, etc.| D[Check comment-style header]
    C -->|skills/udf-*/*.md NEW| E{File type}
    C -->|No match| F[Skip]
    E -->|CUDF_MICROBENCHMARKS.md / references/*.md| G[HTML comment SPDX header]
    E -->|SKILL.md| H[YAML front matter metadata.spdx-file-copyright-text]
    G --> I[Action validates header]
    H --> J{Action supports YAML front matter?}
    J -->|Yes| I
    J -->|Unknown| K[Potential CI failure]
    I --> L{Header present and current year?}
    L -->|Yes| M[Pass]
    L -->|No| N[Fail]
    C -->|skills/udf-*/references/*.md| F

Loading

%%{init: {'theme': 'base', 'themeVariables': {"darkMode": true, "background": "#0d1117", "primaryColor": "#21262d", "primaryTextColor": "#e6edf3", "primaryBorderColor": "#8b949e", "lineColor": "#8b949e", "textColor": "#e6edf3", "edgeLabelBackground": "#161b22", "actorBkg": "#21262d", "actorBorder": "#8b949e", "actorTextColor": "#e6edf3", "actorLineColor": "#8b949e", "signalColor": "#8b949e", "signalTextColor": "#e6edf3", "noteBkgColor": "#373320", "noteBorderColor": "#d4a72c", "noteTextColor": "#f0e6c0", "labelBoxBkgColor": "#21262d", "labelBoxBorderColor": "#8b949e", "labelTextColor": "#e6edf3", "loopTextColor": "#e6edf3", "activationBkgColor": "#30363d", "activationBorderColor": "#8b949e"}}}%%
flowchart TD
    A[PR opened / synchronized] --> B[license-header-check workflow]
    B --> C{File matches included_file_patterns?}
    C -->|*.yml, *.scala, *.py, etc.| D[Check comment-style header]
    C -->|skills/udf-*/*.md NEW| E{File type}
    C -->|No match| F[Skip]
    E -->|CUDF_MICROBENCHMARKS.md / references/*.md| G[HTML comment SPDX header]
    E -->|SKILL.md| H[YAML front matter metadata.spdx-file-copyright-text]
    G --> I[Action validates header]
    H --> J{Action supports YAML front matter?}
    J -->|Yes| I
    J -->|Unknown| K[Potential CI failure]
    I --> L{Header present and current year?}
    L -->|Yes| M[Pass]
    L -->|No| N[Fail]
    C -->|skills/udf-*/references/*.md| F

Loading

_{Reviews (1): Last reviewed commit: "Check UDF skill markdown license headers" | Re-trigger Greptile}

[greptile-apps]

Subdirectory markdown files are not matched by this pattern

skills/udf-*/*.md only matches files exactly one level deep inside each udf-* directory — a bare * does not cross path separators. Four existing markdown files (skills/udf-convert-to-cuda/references/JNI_CUDA_GUIDE.md, NATIVE_BUILD_ENV.md, skills/udf-convert-to-cudf/references/RAPIDS_UDF.md, skills/udf-optimize-cudf/references/OPTIMIZATION_PATTERNS.md) all happen to have headers already, so there is no immediate CI failure, but any future markdown file added under a references/ or other nested directory would silently escape the check. Using skills/udf-*/**/*.md would cover the full tree if that is the intent.

[greptile-apps]

SKILL.md front-matter copyright may not satisfy the license-header action

The SKILL.md files (e.g. skills/udf-benchmark/SKILL.md) embed copyright only in YAML front matter under metadata.spdx-file-copyright-text, not in a comment block. Every non-SKILL markdown file under skills/udf-* (e.g. CUDF_MICROBENCHMARKS.md) uses an <!-- SPDX-FileCopyrightText: … --> HTML comment. If NVIDIA/spark-rapids-common/license-header-check looks for a comment-style header (as it does for .yml, .scala, .py, etc.), all seven SKILL.md files will fail the check the moment any of them is touched in a PR, potentially blocking unrelated work. Has the action been confirmed to accept the YAML front-matter format as a valid copyright declaration? Has the license-header-check action been confirmed (locally or via its source) to accept the metadata.spdx-file-copyright-text YAML front-matter field in SKILL.md files as a valid copyright header, or does it only recognise comment-block formats?