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.
npx skills add lipas-liikuntapaikat/lipas --skill help-video-ingestgit clone --depth 1 https://github.com/lipas-liikuntapaikat/lipasWrote 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/skills/lipas-liikuntapaikat/lipas/help-video-ingest)<a href="https://agentmods.dev/skills/lipas-liikuntapaikat/lipas/help-video-ingest"><img src="https://agentmods.dev/badge/skills/lipas-liikuntapaikat/lipas/help-video-ingest/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/skills/lipas-liikuntapaikat/lipas/help-video-ingest"><img src="https://agentmods.dev/badge/skills/lipas-liikuntapaikat/lipas/help-video-ingest.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00072 | $0.01541 |
| Opus 5 | $0.00036 | $0.00771 |
| Sonnet 5 | $0.00014 | $0.00308 |
| Haiku 4.5 | $0.00007 | $0.00154 |
Grade A, and why
help-video-ingest scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 13d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 147 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Help video ingest
Turn LIPAS tutorial videos (and guide PDFs) into published, translated
help-center pages. The help CMS is the canonical target: published pages
become both human-readable ohjeet and AI-assistant KB docs
(lipas.backend.kb/help-cms->docs — one KB doc per page per language).
All heavy lifting lives in dev/lipas/help_authoring.clj
(lipas.help-authoring, self-contained: loads on any branch, can be
piped inline to remote nREPLs). This skill is the workflow around it.
Prerequisites
GEMINI_API_KEYin the backend JVM env (comes from../.env.sh)- nREPL: local
clj-nrepl-eval -p 7888; remote (e.g. lipas-dev) via an SSH tunnel the user opens, e.g. port 7887 - Load the util locally:
(load-file ".../webapp/dev/lipas/help_authoring.clj")Remote REPLs can'tload-fileyour disk — pipe the file inline:clj-nrepl-eval -p 7887 "$(cat dev/lipas/help_authoring.clj)"
Workflow
1. Analyze the videos
One Gemini call per video, YouTube URL passed natively (fileData.fileUri
— no download; the videos are typically DRM-protected anyway, and agentic
CLIs like agy/gemini can NOT ingest YouTube natively, so don't try
them for this):
(require '[lipas.help-authoring :as ha])
(def material (ha/analyze-video "https://www.youtube.com/watch?v=..."))
(spit "/tmp/extract_<id>.md" material)
~50 s and ~300 input tokens per video-second. The prompt
(ha/extract-prompt) asks for verbatim UI labels, exact mouse/keyboard
mechanics and [MM:SS] timestamps — the value over plain transcripts.
Save each extraction to a file for review; read them all before writing.
The full Lipasinfo video inventory lives in
dev/lipas/kb_ingest.clj (video-sources).
2. Ground against the current product — never skip
Videos age. Verify every UI label and mechanic before it goes into an article:
- Labels:
grep -r "<label>" src/cljc/lipas/i18n/fi/ - Map tool behavior:
src/cljs/lipas/ui/map/(events.cljs,editing.cljs,import.cljs,views.cljs) - When unsure, check the running UI in the browser.
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.
- 13d ago First seen · 147 lines · 72 tokens per session scan A b263d19620d7
help-video-ingest is a skill published in the GitHub repository lipas-liikuntapaikat/lipas (78 stars, last pushed yesterday), licensed MIT. It adds 72 tokens to every session and 1,541 once invoked, about $0.0004 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.
Other skills, from other repositories
new-page
Create a frontend page with routing, i18n, and navigation.
opengis-all
Use when navigating end-to-end GIS workflows — discover the full toolchain from GDAL data processing through GeoServer publishing to CesiumJS/OpenLayers visualization. One-stop index covering the complete open-source GIS data pipeline.
sync-translated-docs
Keep a document and its translated counterpart (README.md and READMEJP.md, docs in two languages) saying the same thing. Use whenever you edit one file of a translated pair, and when asked to check or merge the differences between them. Does not apply to a document that has no counterpart -- a repository with only a…
clojure-expert
Expert knowledge in Clojure/ClojureScript, immutability, REPL-driven development, core.async, and functional programming on the JVM. Use when the user mentions ClojureScript, functional programming, immutability, REPL, core async, or JVM, or when the task involves Lisp Features, Concurrency, REPL-Driven Development…
localize
· Localize multilingual apps: i18n/l10n, translate app strings, add languages, and audit missing translations.
joyride
Joyride core — REPL evaluation, async patterns, VS Code API access, Flares, JS interop, and available libraries. Use when: working with Joyride evaluation, writing ClojureScript in any Joyride context, creating Flares/WebViews, exploring VS Code APIs via the REPL, or using joyrideevaluatecode.