CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

databricks-unstructured-pdf-generation

Build RAG / unstructured-document evaluation datasets and demo documents (e.g. for a Genie Agent with an attached volume) on Databricks: generate synthetic PDFs locally, upload to Unity Catalog volumes, and pair each document with test questions for retrieval evaluation.

Quality

79%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./plugins/databricks/claude/skills/databricks-unstructured-pdf-generation/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

76%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

A highly actionable, well-structured skill body: concrete commands at every step, real gotchas (dbfs: scheme, -r copy semantics, Genie whole-volume indexing), and a documented bundled script. The gaps are the missing validation/verification step after the batch convert-and-upload operations and mild redundancy in the Bundled Script/CLI Reference sections.

Suggestions

Add an explicit verification step after upload, e.g. "databricks fs ls dbfs:/Volumes/my_catalog/my_schema/raw_data/pdf" and compare the file count against ./raw_data/pdf, with a retry path for failed files.

Add a brief check after Step 2 that the output PDF count matches the HTML count (or surface per-file failures from pdf_generator.py) to create a validate→fix→retry loop for the batch conversion.

Tighten the Bundled Script section (it restates the script path and duplicates the CLI Reference) and confirm the ../databricks-genie-agents reference resolves in the installed skill collection.

DimensionReasoningScore

Conciseness

The body assumes Claude's competence (no PDF/Databricks concept explainers) and stays dense, but the Bundled Script section restates the script path twice and the CLI Reference duplicates the script's own --help output. Minor instances of over-explanation that could be trimmed keep it at 4 rather than 5.

4 / 5

Actionability

Every step ships copy-paste-ready commands ("python <SKILL_ROOT>/scripts/pdf_generator.py convert --input ./raw_data/html --output ./raw_data/pdf", "databricks fs cp -r --overwrite ./raw_data/pdf dbfs:/Volumes/..."), a concrete JSON example for the question file, and CLI flags that match the actual bundled script. Fully executable guidance covering the common cases, matching the 5 anchor.

5 / 5

Workflow Clarity

The 4-step workflow is clearly sequenced with per-step commands, but the batch conversion and volume upload have no validation checkpoints (no post-upload "databricks fs ls" verification, no validate-and-retry loop for failed conversions). Per the rubric guideline, batch operations without validation cap workflow_clarity at 3, even though sequence and troubleshooting are otherwise strong.

3 / 5

Progressive Disclosure

Good structure: the single bundled script is one level deep, clearly signaled in a table with a recreate-from-CLI-Reference fallback, and sections are well organized for a short skill. Minor gaps — the cross-skill link "../databricks-genie-agents/references/create-genie-agent.md" is not present in this bundle, and the CLI Reference duplicates the script's own documentation — keep it at 4 rather than 5.

4 / 5

Total

16

/

20

Passed

Description

75%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A specific, well-scoped description with strong concrete actions and a distinct Databricks niche, written in appropriate third-person voice. Its main weakness is the absence of an explicit 'Use when...' trigger clause, which both caps completeness and leaves a few natural trigger synonyms uncovered.

Suggestions

Add an explicit trigger clause, e.g. "Use when building RAG/retrieval evaluation datasets or demo documents on Databricks, or when preparing synthetic PDFs for a Genie Agent with an attached volume."

Include natural trigger variations users would say, such as ".pdf", "eval dataset", "benchmark", or "unstructured documents", to broaden trigger-term coverage.

Keep the existing concrete action list intact — it is the description's strongest asset.

DimensionReasoningScore

Specificity

"generate synthetic PDFs locally, upload to Unity Catalog volumes, and pair each document with test questions for retrieval evaluation" lists multiple specific concrete actions covering the whole pipeline. It matches the 5 anchor (comprehensive coverage) rather than 4, since no step of the workflow is left out.

5 / 5

Completeness

The 'what' is explicit ("Build RAG / unstructured-document evaluation datasets... generate synthetic PDFs locally, upload to Unity Catalog volumes"), but there is no 'Use when...' clause or equivalent explicit trigger guidance; "(e.g. for a Genie Agent with an attached volume)" is an example, not a trigger. Per the rubric guideline, a missing 'Use when' clause caps completeness at 3.

3 / 5

Trigger Term Quality

Natural terms like "RAG", "evaluation datasets", "Genie Agent", "Databricks", "Unity Catalog volumes", and "retrieval evaluation" give good keyword coverage, but common variations are missing (e.g. ".pdf" extension, "benchmark", "test set"). This matches anchor 4 (good coverage, a few natural terms missing) rather than 5's comprehensive synonym/extension coverage.

4 / 5

Distinctiveness Conflict Risk

"on Databricks", "Unity Catalog volumes", and "Genie Agent" carve out a clear niche (synthetic eval datasets on Databricks) with distinct triggers and minimal conflict risk against generic PDF or document skills. It clearly matches the 5 anchor rather than 4's 'minor overlap risk'.

5 / 5

Total

17

/

20

Passed

Validation

87%

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 14 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

frontmatter_unknown_keys

Unknown frontmatter key(s) found; consider removing or moving to metadata

Warning

relative_links

Relative link issues: 1 suspicious

Warning

Total

14

/

16

Passed

Repository

databricks/databricks-agent-skills

Path

plugins/databricks/claude/skills/databricks-unstructured-pdf-generation/SKILL.md

Reviewed

Table of Contents

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.