Skip to content
BeClaude

agent-skill-verification-template

New
1GitHub TrendingGeneralby HelloThisWorld

Production-oriented template for building AI agent skills as verifiable software components — offline eval harness, source-grounding validators, structured logs/traces/metrics, replay artifacts, and a CI quality gate. Runs fully offline with a deterministic mock model.

Summary

This skill provides a production-ready template for building AI agent skills as verifiable software components.

  • It includes an offline eval harness, source-grounding validators, structured logs/traces/metrics, replay artifacts, and a CI quality gate, enabling developers to test and validate agent behavior deterministically without external dependencies.

Install & Usage

1
Create the agents directory
mkdir -p .claude/agents
2
Save the agent file

Add the configuration to .claude/agents/agent-skill-verification-template.md

3
Invoke with @agent-name
@agent-skill-verification-template

Use Cases

Validate that a new agent skill correctly handles edge cases using the offline eval harness before deployment.
Run a CI quality gate to ensure all agent skills meet source-grounding and performance thresholds on every pull request.
Debug a failing agent interaction by replaying structured logs and traces from a previous run.
Benchmark multiple versions of an agent skill against a fixed test suite to compare accuracy and latency.
Generate replay artifacts for compliance or audit purposes, capturing every agent action and decision.
Integrate the deterministic mock model into development workflows to test agent skills without live API costs.

Usage Examples

1

/agent-skill-verification-template run-eval --skill my-skill --test-suite ./tests/eval_suite.json

2

Run the offline eval harness for the customer-support skill with the default mock model and output structured logs.

3

/agent-skill-verification-template ci-gate --skill my-skill --threshold 0.95

View source on GitHub
agent

Security Audits

LicenseUnknownSourceWarnRepositoryPass

Frequently Asked Questions

What is agent-skill-verification-template?

This skill provides a production-ready template for building AI agent skills as verifiable software components. It includes an offline eval harness, source-grounding validators, structured logs/traces/metrics, replay artifacts, and a CI quality gate, enabling developers to test and validate agent behavior deterministically without external dependencies.

How to install agent-skill-verification-template?

To install agent-skill-verification-template: create the agents directory (mkdir -p .claude/agents), then add the config to .claude/agents/agent-skill-verification-template.md. Finally, @agent-skill-verification-template in Claude Code.

What is agent-skill-verification-template best for?

agent-skill-verification-template is a agent categorized under General. It is designed for: agent. Created by HelloThisWorld.

What can I use agent-skill-verification-template for?

agent-skill-verification-template is useful for: Validate that a new agent skill correctly handles edge cases using the offline eval harness before deployment.; Run a CI quality gate to ensure all agent skills meet source-grounding and performance thresholds on every pull request.; Debug a failing agent interaction by replaying structured logs and traces from a previous run.; Benchmark multiple versions of an agent skill against a fixed test suite to compare accuracy and latency.; Generate replay artifacts for compliance or audit purposes, capturing every agent action and decision.; Integrate the deterministic mock model into development workflows to test agent skills without live API costs..