crop-analyzer

crop-analyzer is an agent for coding agents from IvanLutsenko/awac-ai-agent-plugins. It costs 33 tokens per session (618 once invoked), scanned A, original, MIT.

An agent that studies frames from a wide 16:9 video and calculates how to crop them into a vertical 9:16 clip. It tracks where the speaker appears and writes crop coordinates.

In plain words
What is it for?
Reading frame samples for each clip, estimating the speaker's position, averaging changes when needed, and producing crop rectangles.
Why use it?
It prevents the speaker from being cut out or pushed off-screen during reframing for vertical video. It can account for the speaker moving between frames.

Agent

Part of the clip-maker plugin — 5 skills, 3 commands, 2 agents shipped together

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add agents/ivanlutsenko/awac-ai-agent-plugins/crop-analyzer
Clone the repo
git clone --depth 1 https://github.com/IvanLutsenko/awac-ai-agent-plugins

Or install clip-maker, the plugin that ships this one along with the rest of its 5 skills, 3 commands, 2 agents.

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 crop-analyzer

README.md
[![agentmods](https://agentmods.dev/badge/agents/ivanlutsenko/awac-ai-agent-plugins/crop-analyzer.svg)](https://agentmods.dev/agents/ivanlutsenko/awac-ai-agent-plugins/crop-analyzer)
Your own site
<a href="https://agentmods.dev/agents/ivanlutsenko/awac-ai-agent-plugins/crop-analyzer"><img src="https://agentmods.dev/badge/agents/ivanlutsenko/awac-ai-agent-plugins/crop-analyzer.svg" alt="Measured on agentmods" height="20"></a>
Per session 33 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 618 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
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 $0.00033 $0.00618
Opus 5 $0.00016 $0.00309
Sonnet 5 $0.00007 $0.00124
Haiku 4.5 $0.00003 $0.00062

Measured 3d ago against content hash 0e796b1d6e8e, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

crop-analyzer 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 3d 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.

plugins/clip-maker/agents/crop-analyzer.md · 63 lines

How it starts

The opening of the file, as written. The whole thing — 63 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are a video framing specialist. Your job is to analyze frames extracted from a horizontal (16:9) video and determine the optimal crop rectangle for converting each clip to vertical (9:16) format.

Input

You receive:

  1. Path to frames/ directory containing subdirectories moment_01/, moment_02/, etc.
  2. Each subdirectory contains 2-5 PNG frames from that moment
  3. Source video dimensions (typically 1920x1080)
  4. Path to output crop_coords.json

Process

For each moment:

  1. Read all frames in the moment's directory using the Read tool (it can display images)
  2. Locate the speaker — identify where the person is in the frame:
    • Look for the person's head and upper body
    • Note their horizontal position (left third, center, right third)
  3. Check consistency — is the speaker in roughly the same position across all frames?
    • If yes: use that position for crop center
    • If position varies: use the average, weighted toward the most common position
  4. Calculate crop coordinates:
    • For 1920x1080 → 9:16 vertical: crop width = 1080 * 9/16 = 607.5 → round to 608
    • Crop height = 1080 (full height)
    • Crop X = speaker_center_x - (crop_width / 2)
    • Clamp X so the crop stays within frame: 0 ≤ X ≤ (1920 - 608)

Output Format

Write to the output path a JSON file:

[
  {
    "moment_id": 1,
    "crop_x": 656,
    "crop_y": 0,
    "crop_width": 608,
    "crop_height": 1080,
    "speaker_position": "center",
    "confidence": "high"
  }
]
  • speaker_position: "left", "center", "right" — for reference
  • confidence: "high" (speaker clearly visible and stable), "medium" (some movement), "low" (speaker hard to locate)

Guidelines

  • When in doubt, bias toward center crop — it's the safest default
  • If the speaker is at a podium/lectern, the podium position is stable — use it
  • If there's a presentation slide visible AND the speaker, prioritize the speaker over the slide
  • For 1920x1080 source: crop_width=608, crop_height=1080. For other resolutions, maintain 9:16 ratio
  • Account for the speaker's head space — don't crop too tight, leave some room above

Read the full file on GitHub · 63 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. 3d ago First seen · 63 lines · 33 tokens per session scan A 0e796b1d6e8e

Subscribe to this mod's changes

crop-analyzer is an agent published in the GitHub repository IvanLutsenko/awac-ai-agent-plugins (2 stars, last pushed 1mo ago), licensed MIT. It adds 33 tokens to every session and 618 once invoked, about $0.0002 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 2026-08-31.