video-context-grounding

video-context-grounding is a skill for Claude Code, Codex from UCSC-VLAA/VisualClaw. It costs 27 tokens per session (260 once invoked), scanned A, original, MIT.

A video-question method for basing an answer on specific actions and objects that are visibly shown.

In plain words
What is it for?
It is for answering video comprehension questions by linking the chosen option to observed actions, objects, and repeated events.
Why use it?
It prevents plausible-sounding choices from being selected without evidence in the video.

Skill for Claude CodeCodex

Written for no agent in particular: nothing here depends on one.

Good fit It is for answering video comprehension questions by linking the chosen option to observed actions, objects, and repeated events.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/ucsc-vlaa/visualclaw/video-context-grounding
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.

Any agent
npx skills add UCSC-VLAA/VisualClaw --skill video-context-grounding
Clone the repo
git clone --depth 1 https://github.com/UCSC-VLAA/VisualClaw

Made for: Claude Code, Codex.

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 video-context-grounding

README.md
[![agentmods](https://agentmods.dev/badge/skills/ucsc-vlaa/visualclaw/video-context-grounding.svg)](https://agentmods.dev/skills/ucsc-vlaa/visualclaw/video-context-grounding)
Your own site
<a href="https://agentmods.dev/skills/ucsc-vlaa/visualclaw/video-context-grounding"><img src="https://agentmods.dev/badge/skills/ucsc-vlaa/visualclaw/video-context-grounding.svg" alt="Measured on agentmods" height="20"></a>
Per session 27 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 260 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 $0.00027 $0.00260
Opus 5 $0.00014 $0.00130
Sonnet 5 $0.00005 $0.00052
Haiku 4.5 $0.00003 $0.00026

Measured 8d ago against content hash f77540b2a161, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-07, from the pricing page.

Security

Grade A, and why

video-context-grounding 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 8d 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.

visualclaw/skills_seed/seed_universal_mc/video-context-grounding/SKILL.md · 22 lines

What it actually says

Ground Answers in Video Evidence

  1. Before selecting any multiple-choice answer, identify what you actually observe happening in the video (specific actions, objects, sequences).
  2. Match the observed evidence directly to answer options—look for exact correspondence, not paraphrasing or inference.
  3. Reject options that describe actions you did not see, even if they sound plausible or well-written.
  4. For questions about "repetitive actions" or "key developments," explicitly list which actions appeared multiple times or were central to the narrative.
  5. State your reasoning: "The video shows [specific action]. This matches option [X] because [direct evidence]."

Example: Question: What repetitive actions were important? Observed: Sewing occurred twice; cutting occurred once. Correct match: Option (E) mentions both sewing and cutting as repeated/important. Incorrect match: Option (A) mentions only alignment—observed only once, not repetitive.

Anti-pattern: Selecting an answer based on grammatical quality, length, or general plausibility without verifying it matches what is actually visible in the video.

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. 8d ago First seen · 22 lines · 27 tokens per session scan A f77540b2a161

Subscribe to this mod's changes

video-context-grounding is a skill published in the GitHub repository UCSC-VLAA/VisualClaw (55 stars, last pushed 2mo ago), licensed MIT. It adds 27 tokens to every session and 260 once invoked, about $0.0001 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.