Productivity

AI Delivery Evidence Checker

by BEIKE YAN

Overview

Check whether AI coding completion claims are supported by the delivery materials and test evidence you provide. Extract verifiable criteria, identify supported and unsupported work, find evidence gaps, and generate a structured delivery verdict—all without accessing repositories or modifying code.

Tools

check_delivery_evidence

ChatGPT

Use this tool when the user wants to compare structured delivery criteria with a supplied AI completion claim and supplied delivery materials. It checks whether the claim, delivery report, modified-file summaries, diff summary, commit information, and deployment description provide evidence for each criterion. It does not access repositories, verify evidence authenticity, inspect code, run tests, modify files, create commits, deploy software, or guarantee correctness.

check_test_evidence

ChatGPT

Use this tool when the user wants to check whether supplied test summaries, test cases, test results, or independent test descriptions support the relevant delivery criteria and completion claim. It identifies covered, partially covered, and uncovered criteria and records failed, skipped, or missing test scenarios. It does not run tests, access CI systems, inspect code, generate a correctness guarantee, or approve deployment.

extract_delivery_criteria

ChatGPT

Use this tool when the user wants to turn supplied AI coding task requirements into structured delivery criteria. It accepts the original requirements text, optional candidate criteria, and an optional source label, then returns normalized criteria without inventing new requirements. Do not use it to write code, design new requirements, inspect a repository, run tests, or decide whether the delivery is complete.

generate_delivery_verdict

ChatGPT

Use this tool when the user wants a combined evidence-level and delivery-material verdict from structured criteria, delivery-evidence results, test-evidence results, and optional independent verification materials. It returns a material-sufficiency status, blockers, limitations, and rework recommendations. It does not rerun the other tools, inspect code, verify evidence authenticity, approve release, sign off a project, or guarantee that the software is correct or bug-free.

Example Prompts

Click any prompt to copy it.

Links

Documentation ↗Privacy Policy ↗

App Stats

Tools

Prompts

ChatGPT

Platforms