deepeval-tracing

Instrument an AI application with DeepEval's native tracing so its behavior is visible in Confident AI. TRIGGER when the user wants to add DeepEval tracing or @observe to an LLM app, agent, RAG pipeline, or chatbot; wire a framework, model-provider, or vector-database integration (LangGraph, LangChain, OpenAI Agents, LlamaIndex, Pydantic AI, CrewAI, and others); choose between a native integration and manual instrumentation; set span types, tags, or metadata; or send DeepEval-SDK traces to Confident AI's Observatory. DO NOT TRIGGER for building DeepEval pytest eval suites, datasets, goldens, metrics, or deepeval test run (use the `deepeval` skill), or for raw OpenTelemetry / OTLP export without the deepeval package (use the `deepeval-otel` skill). This skill is purely DeepEval-SDK instrumentation — producing well-formed traces, not running evals.

Quality

89%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Quality

Content

78%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.

Well-structured, lean, and navigable body that orients the reader and defers detail to one-level-deep references. The main improvements are de-duplicating the repeated scope/boundary statements and adding an inline @observe example plus a verification remediation loop.

Suggestions

Consolidate the 'instrument AI components only' and 'deepeval vs deepeval-otel boundary' statements so each appears once (e.g. in Scope and When-to-Use) rather than echoing through Core Principles, reducing token cost.

Add a minimal inline `@observe` code snippet in the Workflow or a short Examples section so the core instrumentation pattern is copy-paste ready without opening references.

Expand step 7 into a feedback loop: if traces do not appear in the Observatory, check `deepeval login`/`CONFIDENT_API_KEY` and re-verify before finishing.

Dimension	Reasoning	Score
Conciseness	The body is lean and avoids explaining concepts Claude already knows, but the 'AI components only' scope and the skill-boundary distinction (vs deepeval/deepeval-otel) are restated across the intro, Scope, When-to-Use, and Core Principles sections — minor trimming possible.	4 / 5
Actionability	Concrete steps, specific span types (llm/retriever/tool/agent), and real commands (pip install deepeval, deepeval login, CONFIDENT_API_KEY) are present, but the core @observe instrumentation code lives only in references rather than as an inline copy-paste example.	4 / 5
Workflow Clarity	A clearly sequenced 7-step workflow with a stop-condition in step 1 and a verification step (confirm traces appear) in step 7, though it lacks an explicit fix-and-retry loop if verification fails.	4 / 5
Progressive Disclosure	SKILL.md is a concise overview pointing to two real one-level-deep reference files (references/tracing.md, references/integrations.md), signaled both inline in the workflow and in a References table for easy navigation.	5 / 5
	Total	17 / 20 Passed

Description

100%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.

An exemplary description: concrete actions, rich natural trigger terms, explicit what-and-when guidance, and clear boundaries against related skills. No meaningful gaps to address.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions — instrumenting with native tracing, wiring framework/model/vector-DB integrations, choosing native vs manual instrumentation, setting span types/tags/metadata, and sending traces to the Observatory — giving comprehensive coverage rather than vague language.	5 / 5
Completeness	Explicitly states what the skill does (instrument an AI app with DeepEval native tracing visible in Confident AI) and when to use it via concrete 'TRIGGER when…' and 'DO NOT TRIGGER for…' clauses.	5 / 5
Trigger Term Quality	Comprehensive natural-term coverage including '@observe', 'LLM app, agent, RAG pipeline, or chatbot', explicit framework names (LangGraph, LangChain, OpenAI Agents, LlamaIndex, Pydantic AI, CrewAI), and synonyms a user would actually say; uses third-person voice.	5 / 5
Distinctiveness Conflict Risk	Carves a clear niche (DeepEval-SDK instrumentation only) and explicitly disambiguates against the sibling `deepeval` and `deepeval-otel` skills, minimizing wrong-skill triggering.	5 / 5
	Total	20 / 20 Passed

Validation

100%

Warnings & errors only

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

Validation — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: confident-ai/deepeval
Path: skills/deepeval-tracing/SKILL.md
Commit: 11e53e8

Reviewed: about 14 hours ago

Table of Contents

Discovery Implementation Validation

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.