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 parijatmukherjee/openclaw-hawkins --skill vision-agent-skillgit clone --depth 1 https://github.com/parijatmukherjee/openclaw-hawkinsWrote 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/parijatmukherjee/openclaw-hawkins/vision-agent-skill)<a href="https://agentmods.dev/skills/parijatmukherjee/openclaw-hawkins/vision-agent-skill"><img src="https://agentmods.dev/badge/skills/parijatmukherjee/openclaw-hawkins/vision-agent-skill/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/parijatmukherjee/openclaw-hawkins/vision-agent-skill"><img src="https://agentmods.dev/badge/skills/parijatmukherjee/openclaw-hawkins/vision-agent-skill.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 | $0.00043 | $0.00536 |
| Opus 5.5 | $0.00017 | $0.00214 |
| Sonnet 5.5 | $0.00009 | $0.00107 |
| Haiku 4.5 | $0.00004 | $0.00054 |
Grade A, and why
vision-agent-skill 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 — 74 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Vision Agent
Purpose
Specialized agent for tasks requiring image understanding and visual analysis.
When to Use
Spawn this agent when the task involves:
- Analyzing screenshots or images
- Reading text from images (OCR)
- Interpreting charts, graphs, or diagrams
- Visual debugging (UI layouts, error screenshots)
- Image comparison or quality assessment
- Any task where "seeing" the image matters
Model
- Primary:
ollama/kimi-k2.5:cloud(text+image, 125k context) - Why: kimi-k2.5 is vision-capable via Ollama cloud. It can accept image inputs and analyze them.
Data handling (read first)
Images and screenshots are sent to a cloud-hosted vision model — they leave this host. Screenshots in particular routinely contain sensitive data: credentials or tokens on screen, API keys, personal/customer information, internal URLs and hostnames, or private messages. Before sending an image:
- Confirm the operator is OK with this specific image going to the cloud model.
- Prefer redacting or cropping out anything sensitive that isn't needed for the task.
- Never upload an image solely to extract a secret from it.
How to Spawn
sessions_spawn(
task="[VISION] Analyze this screenshot and tell me what's wrong with the layout",
mode="run",
model="ollama/kimi-k2.5:cloud",
timeoutSeconds=300
)
Task Prefix
Always prefix vision tasks with [VISION] so the Nexus routes them correctly:
[VISION]— Image analysis, screenshot reading, visual tasks → routes tovision-agent
Examples
# Analyze a screenshot
sessions_spawn(
task="[VISION] Look at this screenshot and identify any UI errors or layout issues",
mode="run"
)
# Read text from image
sessions_spawn(
task="[VISION] Extract all text from this image and format it as markdown",
mode="run"
)
# Compare images
sessions_spawn(
task="[VISION] Compare these two screenshots and list the differences",
mode="run"
)
Communication
- The Nexus stays conversational
- Acknowledge before spawning: "I'll delegate this image analysis to the vision agent."
- Synthesize results when complete
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 · 74 lines · 43 tokens per session scan A 86a04c261de6
vision-agent-skill is a skill published in the GitHub repository parijatmukherjee/openclaw-hawkins (5 stars, last pushed 4mo ago), licensed MIT. It adds 43 tokens to every session and 536 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 2026-08-31.
Other skills, from other repositories
openhermit-admin
Explain what OpenHermit is and how to deploy and administer it — setup, CLI, gateway, agent config, API. Use when the user asks about installing, operating, or administering an OpenHermit deployment. For how agents behave for their owners/users/guests, see openhermit-guide.
skill-creator
Create, structure, and validate new skills. Use when designing a skill, scaffolding a skill directory, or improving an existing skill.
openhermit-guide
Explain how an OpenHermit agent behaves for the people interacting with it — owners, users, and guests. Covers roles, identity linking, the agent's tools, access levels, channels, the policy model, and the approval flow. Use when someone asks how an OpenHermit agent works from a caller's perspective. For deploying or…
github
Manage GitHub issues and pull requests, inspect CI runs, and query repository data.
mission-control
Interact with Mission Control — AI agent orchestration dashboard. Use when registering agents, managing tasks, syncing skills, or querying agent/task status via MC APIs.
skill-creator
Generates Anthropic Skills with complete workflow including GitHub PR creation and local download verification.