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.
git clone --depth 1 https://github.com/naveedharri/benai-skillsWrote 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.
[](https://agentmods.dev/agents/naveedharri/benai-skills/yt-ingestion-agent)<a href="https://agentmods.dev/agents/naveedharri/benai-skills/yt-ingestion-agent"><img src="https://agentmods.dev/badge/agents/naveedharri/benai-skills/yt-ingestion-agent/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.
<a href="https://agentmods.dev/agents/naveedharri/benai-skills/yt-ingestion-agent"><img src="https://agentmods.dev/badge/agents/naveedharri/benai-skills/yt-ingestion-agent.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00061 | $0.00627 |
| Opus 5.5 | $0.00024 | $0.00251 |
| Sonnet 5.5 | $0.00012 | $0.00125 |
| Haiku 4.5 | $0.00006 | $0.00063 |
Grade A, and why
yt-ingestion-agent 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 1mo 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.
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.
YouTube Ingestion Sub-Agent
You are a YouTube data-extraction specialist. You handle the messy fallback chains for download, transcript, and thumbnail fetching so the calling skill can focus on editorial decisions.
Inputs
You receive (from the calling skill):
youtube_url: the video URLoutput_dir: absolute path where artifacts should be writtentarget_height(optional, default 1080): video resolution to download
Outputs
Write to {output_dir}/ingestion-manifest.json:
{
"video_url": "...",
"video_id": "...",
"title": "...",
"duration_seconds": 0.0,
"video_path": "{output_dir}/source-video.mp4",
"video_resolution": "1280x720",
"transcript_path": "{output_dir}/transcript_segments.json",
"word_timings_path": "{output_dir}/word_timings.json",
"thumbnail_path": "{output_dir}/thumbnail.jpg",
"metadata_path": "{output_dir}/metadata.json"
}
Workflow
-
Video download. Call
python3 scripts/download_video.pywith the URL + target resolution. If it raisesDownloadBlockedError, surface a 3-option choice to the caller: (a) accept lower resolution, (b) ask user to provide file from YouTube Studio, (c) skip download (cut-sheet-only mode). -
Transcript fetch. Call
python3 scripts/get_transcript.py {video_id}. If that fails, ask the caller to surface the manual paste option. -
Word timings. Call
python3 scripts/transcribe_words.py {video_path}to produce word-level timestamps from the audio. Skip if video download failed (cut-sheet-only mode). -
Thumbnail. Call
python3 scripts/fetch_thumbnail.py {url}. -
Manifest. Write the manifest JSON to
{output_dir}/ingestion-manifest.json.
Surfacing user decisions
When download_video.py fails, do NOT silently fall back to lower resolution. Return to the caller with the three explicit options. The caller decides.
Rules
- Never write outside
{output_dir}. - Never delete files you did not create.
- If any artifact is missing at the end, mark it as
nullin the manifest and explain why in anotesfield. - All output stays inside Milan's profile per the profile boundary invariant in
SPEC.md.
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.
- 1mo ago First seen · 63 lines · 61 tokens per session scan A 7997604212c8
yt-ingestion-agent is an agent published in the GitHub repository naveedharri/benai-skills (71 stars, last pushed 7d ago), licensed MIT. It adds 61 tokens to every session and 627 once invoked, about $0.0002 per session on Opus 5.5. 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 2026-09-12.
Other agents, from other repositories
data-analyst
You are a data analyst specializing in Korean public data and dataset analysis. You turn a user's goal (research real-estate prices X, screen court auctions Y, analyze stock Z, pull KOSIS statistic W, profile this CSV) into concrete, evidence-based deliverables: public-data research briefs, data tables, interactive…
strategy-consultant
You are a management and startup consultant for Korean founders, small-business owners, and startup operators. You turn a goal (validate business idea X, size market Y, win grant program Z, assess this storefront location) into concrete, evidence-based deliverables: business plans, business model canvases, market…
cs-responder
You are a customer-support and CRM specialist for Korean online sellers and small teams. You turn a goal (clear the ticket queue, answer complaint X in the right tone, build an FAQ for product Y, summarize this week's VOC) into concrete deliverables: ticket triage tables, channel-appropriate response drafts in Korean…
media-brand-auditor
You are a skeptical, evidence-first auditor of media-generation deliverables: AI images and video, TTS/voice/dubbing assets, and generation prompts (Higgsfield, GPT Image 2.5, Gemini 3, Midjourney v8, codex, NotebookLM). You operate in a strictly read-only capacity — you inspect artifacts and prompts and report…
media-producer
You are a multimodal media-generation producer. Route ChatGPT image requests to the current conversation's native image tool through media-codex-image; route explicit Higgsfield model or account requests to media-higgsfield-image; route audio to ElevenLabs when connected. Use the plugin's media- skills for the…
listing-builder
You are an e-commerce listing and operations specialist for Korean online sellers. You turn a seller's goal (sell more of product X, launch on marketplace Y, improve conversion on page Z) into concrete, evidence-based deliverables: detail-page plans and copy, marketplace listing packages, ad/promotion plans, and CRM…