.claude/skills/cercano-classify/SKILL.md
Classify or triage text using local AI via Cercano. Returns a category, confidence level, and brief reasoning. Use this for quick local triage of errors, logs, code quality issues, bug reports, or any content that needs categorization without sending it to the cloud.
npx skillsauth add GoogleDevRelExplorations/cercano cercano-classifyInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
Classify or triage text into categories using local AI inference.
Tool name: cercano_classify
| Parameter | Type | Required | Description |
|-----------|------|----------|-------------|
| text | string | Yes | The text to classify or triage. |
| categories | string | No | Comma-separated list of categories to choose from. If omitted, the model determines appropriate categories. |
Returns a structured classification:
Classify an error with predefined categories:
{
"text": "FATAL: password authentication failed for user \"postgres\"",
"categories": "auth,network,configuration,data,unknown"
}
Open-ended classification:
{
"text": "The API returns 200 but the response body is empty when the user has no profile picture set"
}
devops
Summarize text or files using local AI via Cercano without sending content to the cloud. Supports brief, medium, and detailed summary lengths.
devops
Submit cloud token usage data to Cercano (opt-in). This sends data, not a report — use cercano_stats to view usage. Helps track cloud tokens alongside local inference for accurate local-vs-cloud comparison.
tools
View Cercano usage statistics and cloud token savings. Shows total requests, tokens processed locally, cloud tokens reported by the host, percentage kept local, and breakdowns by tool, model, and day.
development
Research a question using DuckDuckGo search and local AI analysis. Crafts search queries, fetches top results, and synthesizes a sourced answer — all locally.