CtrlK
BlogDocsLog inGet started
Tessl Logo

snowpipe-streaming-ai-webinar

End-to-end Snowpipe Streaming HPA + AI demo for webinars. Sets up Snowpipe Streaming high-performance architecture (HPA) with background data generation, deploys a live Streamlit dashboard, then layers on a Semantic View and Cortex Agent so the presenter can do natural-language queries on live-streaming data in Snowsight. Triggers: snowpipe streaming ai webinar, snowpipe streaming webinar demo, snowpipe streaming ai demo, streaming ai demo, snowpipe streaming webinar, snowpipe streaming cortex agent demo, snowpipe streaming semantic view demo.

80

1.20x
Quality

83%

Does it follow best practices?

Impact

100%

1.20x

Average score across 2 eval scenarios

SecuritybySnyk

High

Do not use without reviewing

SKILL.md
Quality
Evals
Security

Quality

Content

70%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-sequenced runbook: every step has concrete commands or SQL, explicit stop conditions, bounded retries, and user confirmation before destructive cleanup. Its weaknesses are token efficiency (heavy internal repetition and a fully inlined 125-line script) and progressive disclosure — the missing Streamlit template source referenced by Step 5a is a genuine defect, and the large code blocks belong in reference files.

Suggestions

Add the full Streamlit app source to the Templates section (or a references/streamlit_app.py.md file) — Step 5a's fallback path currently points to content that does not exist in the skill.

Deduplicate: state the 30-minute background-streaming rule and the batching rules once, and drop the redundant table DDL from the Templates section since it already appears in Step 3.

Move the full streaming_demo.py and the semantic-view CREATE statement into a references/ file, keeping only key configuration differences inline, to cut the ~800-line body down toward the core workflow.

DimensionReasoningScore

Conciseness

The body is instruction-dense rather than explanatory, but it is noticeably repetitive and could be tightened: the 'Do NOT ask about demo duration' instruction appears three times (execution guidelines, Step 1b, and the streaming script comment), the batching rules are restated in both the guidelines and Step 5, the full table DDL is duplicated verbatim in the Templates section, and a ~125-line Python script is inlined in full. It fits 'mostly efficient but includes some unnecessary explanation or could be tightened' rather than the lean 5 or the padded 2.

3 / 5

Actionability

Guidance is overwhelmingly concrete and executable — exact bash commands, complete SQL blocks, a full copy-paste-ready streaming script with placeholders, and precise parallel-tool-call instructions. The gap: Step 5a says 'fall back to writing the full Streamlit app inline (see the Templates section at the end of this skill for the full source)', but the Templates section contains only the table DDL and pipe convention — the promised Streamlit source is absent, leaving one path non-executable as written. This fits 'mostly executable guidance with minor gaps' rather than fully copy-paste-ready 5.

4 / 5

Workflow Clarity

Steps 0-10 are clearly sequenced with explicit validation checkpoints throughout: Step 1 lists hard stop conditions (Python <3.9, glibc <2.26, NULL warehouse), Step 3 confirms the RSA key via DESC USER, Step 6a retries row_count with a bounded retry loop before proceeding, SHOW statements verify each created object, and the destructive Step 10 cleanup asks the user to choose an option before dropping anything. This matches the anchor for clear sequencing, explicit validation, and feedback loops.

5 / 5

Progressive Disclosure

The step-based structure with a 'What this skill provides' summary is navigable, but no bundle files exist (references/, scripts/, assets/ are absent), and content that would naturally live in separate files — the full streaming demo script, the semantic-view SQL, and the promised Streamlit app source — is inlined in a ~800-line body. Worse, the one reference it does make (the 'Templates section' for the Streamlit source) points to content that is not there. This fits 'some structure but could be better organized' rather than the well-split 4 or 5.

3 / 5

Total

15

/

20

Passed

Description

96%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 strong description: third-person voice, dense with concrete actions, and an explicit multi-phrase Triggers clause that clearly answers both what the skill does and when to use it. The only weakness is mild overlap with the closely related quickstart and semantic-view/cortex-agent skills, which the body (but not the description) disambiguates.

DimensionReasoningScore

Specificity

The description lists multiple concrete actions — 'Sets up Snowpipe Streaming high-performance architecture (HPA) with background data generation, deploys a live Streamlit dashboard, then layers on a Semantic View and Cortex Agent' — covering the full pipeline comprehensively. It matches the anchor for multiple specific concrete actions with comprehensive coverage; score 4's 'minor gaps in coverage' does not apply since setup, dashboard, semantic view, agent, and NL querying are all named.

5 / 5

Completeness

It explicitly answers both questions: 'what' via the concrete action list (Snowpipe Streaming HPA setup, background data generation, Streamlit dashboard, Semantic View, Cortex Agent, natural-language queries in Snowsight) and 'when' via the explicit 'Triggers:' clause with concrete phrases plus 'for webinars'. This matches the anchor for clearly and explicitly answering both what AND when with concrete trigger phrases.

5 / 5

Trigger Term Quality

The 'Triggers:' clause provides comprehensive natural-phrase coverage with many variations users would actually say: 'snowpipe streaming ai webinar', 'snowpipe streaming webinar demo', 'snowpipe streaming ai demo', 'streaming ai demo', 'snowpipe streaming cortex agent demo'. These are the natural phrasings for this need, matching the anchor for comprehensive coverage including synonyms; nothing relevant is missing for this niche.

5 / 5

Distinctiveness Conflict Risk

The niche is clear — a presentation-ready webinar demo combining streaming + AI — and all triggers are prefixed with 'snowpipe streaming', keeping conflict risk low. However, there is minor overlap risk with the closely related 'snowpipe-streaming-quickstart' skill and bundled 'semantic-view'/'cortex-agent' skills, since triggers like 'snowpipe streaming webinar demo' describe a superset of the quickstart's domain. This fits 'mostly distinct; minor overlap risk with closely related skills' better than the clear-niche 5 anchor.

4 / 5

Total

19

/

20

Passed

Validation

93%

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

Validation — 15 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

skill_md_line_count

SKILL.md is long (800 lines); consider splitting into references/ and linking

Warning

Total

15

/

16

Passed

Repository
snowflakedb/snowpipe-streaming-sdk-examples
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.