davinci-resolve-mcp: Agent for Claude Code

.claude/agents/grade-match-verifier.md

grade-match-verifier is an agent for Claude Code from samuelgursky/davinci-resolve-mcp. It costs 63 tokens per session (745 once invoked), scanned A, original, MIT.

A measurement-based check for whether graded video shots visually match. It compares rendered frames from DaVinci Resolve using color and brightness measurements instead of relying on metadata or visual guesses.

In plain words
What is it for?
Use it after matching shots or applying a grade to verify pairs of shots, calculate color and brightness differences, and receive a pass or fail against the project's tolerance.
Why use it?
Two shots can have the same grade settings but still look different because their original exposure or white balance differed.

Agent for Claude Code

Written for Claude Code: installed under .claude/. Also seen: model in frontmatter.

This is samuelgursky/davinci-resolve-mcp's own configuration. It tells Claude Code how to work on davinci-resolve-mcp itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything davinci-resolve-mcp configures →

About the project

davinci-resolve-mcp is an MCP server that lets AI assistants control DaVinci Resolve Studio through its scripting API, including editing, media organization, rendering, grading, and other project tasks. It is for users who want agents to operate Resolve and inspect media through a local control panel. The catalogue entries support agent workflows for this Resolve integration.

samuelgursky/davinci-resolve-mcp · 2,622 stars · on GitHub

Reuse

Borrowing it

Nothing to install: this file belongs to samuelgursky/davinci-resolve-mcp. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file
curl -O https://raw.githubusercontent.com/samuelgursky/davinci-resolve-mcp/main/.claude/agents/grade-match-verifier.md
Clone the repo
git clone --depth 1 https://github.com/samuelgursky/davinci-resolve-mcp

Made for: Claude Code.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for grade-match-verifier

README.md
[![agentmods](https://agentmods.dev/badge/agents/samuelgursky/davinci-resolve-mcp/grade-match-verifier/github.svg)](https://agentmods.dev/agents/samuelgursky/davinci-resolve-mcp/grade-match-verifier)
Your own site
<a href="https://agentmods.dev/agents/samuelgursky/davinci-resolve-mcp/grade-match-verifier"><img src="https://agentmods.dev/badge/agents/samuelgursky/davinci-resolve-mcp/grade-match-verifier/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for grade-match-verifier

Your own site · 80×15
<a href="https://agentmods.dev/agents/samuelgursky/davinci-resolve-mcp/grade-match-verifier"><img src="https://agentmods.dev/badge/agents/samuelgursky/davinci-resolve-mcp/grade-match-verifier.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 63 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 745 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00063 $0.00745
Opus 5 $0.00032 $0.00373
Sonnet 5 $0.00013 $0.00149
Haiku 4.5 $0.00006 $0.00075

Measured 11d ago against content hash 91a0a55c85d4, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-10, from the pricing page.

Security

Grade A, and why

grade-match-verifier scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 11d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

.claude/agents/grade-match-verifier.md · 65 lines

How it starts

The opening of the file, as written. The whole thing — 65 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Grade Match Verifier

You answer one question: do these shots match? With numbers, from rendered frames. Not from the node graph, not from CDL values, not from an impression of a thumbnail.

Why this agent exists

"Matched" is the quality bar that gets claimed most and verified least. Two clips can carry identical grade metadata and still not match, because they started from different exposures and white points. The only evidence that counts is measured pixels out of Resolve.

Method

  1. Establish the hero. Whichever shot the match was made to. If the user did not name one, ask the main session rather than guessing — measuring against the wrong reference produces confidently wrong numbers.

  2. Render frames at matched timecodes. Use gallery_stills export or media_analysis frame extraction. Every frame must come out of Resolve with the grade active — a frame pulled from the source file measures the camera, not the grade.

  3. Measure. Sample each frame and compute, per shot:

    • mean R, G, B
    • R−B delta — the primary warm/cool axis this project judges on
    • mean luma, and the shadow/highlight ends separately

    Restrict sampling to comparable content. Do not measure across a whole frame when the shots differ in composition; a bright sky in one and not the other will swamp the result.

    Skin-mask trap: if you mask to skin tones to compare faces, verify the mask actually caught pixels in both frames before trusting the comparison. An empty or near-empty mask returns a delta near zero and reads as a perfect match — this has hidden a real, visible correction before. Report the pixel count behind every masked measurement, and treat any mask under a few hundred pixels as no measurement at all.

  4. Judge against tolerance. The working standard on this project is an R−B delta at or under 0.012 between shots that should match. Report the measured number for every pair, not just pass/fail.

Read the full file on GitHub · 65 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 11d ago First seen · 65 lines · 63 tokens per session scan A 91a0a55c85d4

Subscribe to this mod's changes

grade-match-verifier is an agent published in the GitHub repository samuelgursky/davinci-resolve-mcp (2,622 stars, last pushed today), licensed MIT. It adds 63 tokens to every session and 745 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Related

Other agents, from other repositories

e2e-media-processor

Post-processes E2E artifacts into GIF/MP4/thumbnail (browser screenshots or asciinema .cast). Dispatched by e2e-test, e2e-flow, e2e-walkthrough.

iamcxa/kc-claude-plugins · 50 tokens

finalize

Snapshot visual QA + one in-place fix pass + render. Dispatched only when Step 6 lint/inspect reports issues, or to do the final render.

vibe-motion/auto-motion · 0 tokens

demo-frame-critic

Reviews a rendered demo video by reading extracted frames and grading each section against its successCriteria — finds errors, empty states, spinners, placeholder data, exposed secrets, clipped text, caption collisions, and pacing problems. Returns a ranked findings list with the stage to re-run for each. Use after a…

lukaskellerstein/claude-my-marketplace · 156 tokens

pixel-art-animation-reviewer

Independent reviewer of pixel-art ANIMATION quality (loop seamlessness, motion physics, multi-component motion, frame timing, period selection, particle determinism). One of four specialized review roles in the pixel-art-quality-board orchestrator. Use when the user asks to "check animation timing", "verify loop…

AnastasiyaW/codex-claude-code-config · 140 tokens

e2e-tester

Use for end-to-end and smoke testing of critical user paths across viewports. Pairs with a browser-automation MCP (for example Playwright) when one is available.

mnzralee/claude-multi-agent-architecture · 41 tokens

final-report-reviewer-agent

Auto-Harness reviewer subagent for final QA report compliance. Use only immediately after evaluatorfinal writes the final QA report.

redker56/auto-harness · 30 tokens