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 trrrrrryg/dsh-visual-skin --skill deepseek-harness-skin-studiogit clone --depth 1 https://github.com/trrrrrryg/dsh-visual-skinWrote 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/trrrrrryg/dsh-visual-skin/deepseek-harness-skin-studio)<a href="https://agentmods.dev/skills/trrrrrryg/dsh-visual-skin/deepseek-harness-skin-studio"><img src="https://agentmods.dev/badge/skills/trrrrrryg/dsh-visual-skin/deepseek-harness-skin-studio/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/trrrrrryg/dsh-visual-skin/deepseek-harness-skin-studio"><img src="https://agentmods.dev/badge/skills/trrrrrryg/dsh-visual-skin/deepseek-harness-skin-studio.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 | $0.00098 | $0.00930 |
| Opus 5 | $0.00049 | $0.00465 |
| Sonnet 5 | $0.00020 | $0.00186 |
| Haiku 4.5 | $0.00010 | $0.00093 |
Grade A, and why
deepseek-harness-skin-studio 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 12d 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 — 40 lines — stays where its author put it; the contents beside it link to each section on GitHub.
DeepSeek Harness Skin Studio
Use the project-owned MCP tools when available. Keep all theme state in the shared DesignSession; never maintain a private copy in chat.
Automatic skin creation
When the user asks to create or replace a skin (often with just a background image), drive the whole flow through MCP:
design_create(or load the active design withtheme_getwhen the user wants to modify the current skin).asset_uploadthe provided image (PNG/JPEG/WebP, up to 4 MiB) to get its content-addressedassetId.theme_patchthe design: setappearance.backdrop(andregions.sidebar/regions.mainwhen split mode is wanted) to{ kind: "image", assetId, fit, position, overlay }, plus tasteful token overrides. Always use the currentbaseRevision; onREVISION_CONFLICTreload withtheme_getand retry.theme_validate, thenpreview_startso the user can see the new skin live in Studio (isolated DSH; it never touches the live profile).- Before persisting, present the apply plan and require the human to press Confirm in the visible Studio; after the user-controlled DSH restart, verify plugin health and the acknowledged revision/hash.
In DSH, the tools are exposed as mcp__skin-studio__* (doctor, design_create, asset_upload, theme_patch, theme_get, theme_validate, preview_start, theme_apply_plan, theme_apply, ...); in Codex they appear under the native names. Prefer these tools over manual file edits.
Studio session flow
- Run
scripts/open-studio.ps1immediately. It validates the versioned runtime bundled inside this Skill, starts the Controller when needed, and opens the independent Studio. Report its copyable URL if automatic opening fails. - Run
doctor, then usestudio_openfor subsequent opens in the same session. - Create or load a design, then use
theme_patchwith the currentbaseRevision. OnREVISION_CONFLICT, reload withtheme_get, reconcile, and retry. - Use
theme_validate,preview_start, andpreview_snapshotwhile designing.preview_startcreates a Controller-owned, isolated actual DSH runtime under a temporary home; it never writes to, restarts, or loads the user's DSH profile. Wait for the current revision'sliverender receipt before proposing apply. The user may edit the same revision in the web UI. - Before apply or restore, ask the user to review the immutable apply plan and press Confirm in the visible Studio. The Studio itself performs the confirmed write only when the plan is bound to the current isolated preview session, generation, and render receipt. Never create, simulate, retrieve, transmit, or reuse a confirmation token.
- Agents only observe
operation_status. After the user-controlled DSH restart, verify plugin health and the acknowledged design revision/hash; never report success while the operation is pending verification.
What ships with it
2 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 12d ago First seen · 40 lines · 98 tokens per session scan A d936b52b9b0f
deepseek-harness-skin-studio is a skill published in the GitHub repository trrrrrryg/dsh-visual-skin (1 stars, last pushed 12d ago), licensed MIT. It adds 98 tokens to every session and 930 once invoked, about $0.0005 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.
Other skills, from other repositories
xiaohongshu-cover
An AI cover-planning tool for Xiaohongshu, a Chinese social platform for lifestyle and product content. It studies popular covers in a topic area and produces three cover concepts with examples and image-generation prompts.
dsh-web-skin-developer
Build a new skin for the dsh-web skin collection (DSH Web GUI) and publish it into the Skin Center — the first-level settings section — scaffold with scripts/dsh-skin-new, author the v2 skin.json manifest plus skin.css token remap (pure asset directory, no package.json, no build step), validate with scripts/dsh-skin…
playlet-wechat-feed
A daily report of popular short-drama articles from WeChat public accounts, a Chinese publishing platform, grouped by story themes and accompanied by writing insights.
wechat-creator
A WeChat Official Account content and analysis tool for finding popular articles, creating and rewriting posts, scoring titles, designing covers, checking restricted words, and diagnosing accounts. WeChat Official Accounts are publishing channels inside the WeChat messaging platform.
xiaohongshu-search
A tool for searching and ranking popular Xiaohongshu posts, a Chinese social content platform. It uses keywords, engagement, relevance, and recency to recommend content and related search directions.
douyin-works-crawler
A tool that retrieves basic information and recent videos from a Douyin account, using its name or Douyin ID. It can return up to 50 videos with engagement data and links.