dsh-web is a plugin ecosystem that adds a web interface and optional tools to DeepSeek Harness, an AI development environment. It is used to provide features such as task boards, mobile remote control, SSH operations, image understanding, file and code panels, Git views, agent presets, and interchangeable skins through independently loaded plugins. The catalogue skills are plugins and related workflows for extending and operating the DSH Web environment.
Borrowing it
Nothing to install: this file belongs to zhu1090093659/dsh-web. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/zhu1090093659/dsh-web/dev/.agents/skills/dsh-web-agent-coding/SKILL.mdgit clone --depth 1 https://github.com/zhu1090093659/dsh-webWrote 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/zhu1090093659/dsh-web/dsh-web-agent-coding)<a href="https://agentmods.dev/skills/zhu1090093659/dsh-web/dsh-web-agent-coding"><img src="https://agentmods.dev/badge/skills/zhu1090093659/dsh-web/dsh-web-agent-coding/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/zhu1090093659/dsh-web/dsh-web-agent-coding"><img src="https://agentmods.dev/badge/skills/zhu1090093659/dsh-web/dsh-web-agent-coding.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 | $0.00030 | $0.00935 |
| Opus 5.5 | $0.00012 | $0.00374 |
| Sonnet 5.5 | $0.00006 | $0.00187 |
| Haiku 4.5 | $0.00003 | $0.00093 |
Grade A, and why
dsh-web-agent-coding 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 — 36 lines — stays where its author put it; the contents beside it link to each section on GitHub.
dsh-web Agent Coding
Root instructions own repository safety, branches, and merge gates. This skill owns the working method; it does not grant Git or release authorization.
Scope and implement
- Confirm the root, branch, and dirty baseline before writing; preserve other sessions' files and index. Read only applicable directory instructions: packages for package work, docs for documentation, then the nearest owner.
- Change the smallest owner: a plugin, skin, shared source, generator, or aggregate. Keep host/client/shared boundaries and browser platform-import/type-only SDK contracts in the package instructions. Modify generated copies through their source and generator, not by hand.
- Follow Agent Note rules for non-trivial decisions. Reuse the owning note where appropriate; do not copy its lifecycle or format rules here.
- Verify the affected behavior and review the diff. Run focused checks during edits, and the repository's complete required gates at merge/push/release boundaries, not after every local change. Report actual evidence and limitations; a commit is not delivery.
Context, delegation, and failure recovery
- Load task-relevant references, not the entire repository map. When authoring prompts, keep stable rules separate from changing evidence (queries, diffs, task state); place dynamic material after stable context where supported. Do not rewrite runtime-owned system/tool messages, assume cache hits, or change instruction priority for caching.
- Delegate independent, bounded work only when it helps. Select from available models by task difficulty, error cost, latency, and observed quality, not fixed provider aliases. Use capable reasoning for consequential ambiguity; do not infer that all documentation or static analysis is low-risk. Pass only necessary context and tools where the runtime permits.
- On a failing check, capture the exact log and exit code, reproduce with the narrowest useful command, fix the evidenced cause, and rerun the affected check. Separate pre-existing failures from regressions caused by this change; do not mix unrelated refactors into repairs.
- Stop repeating an approach when it adds no evidence or edits oscillate. State the obstacle, inspect a different cause or choose a materially different safe approach. Ask the user only for missing information, authorization, or a consequential choice; a retry count alone is not a reason to ask. Report an environment blocker rather than bypassing safety or weakening a gate.
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 Changed 851953092552
- 1mo ago Changed c198cf3f77f3
- 1mo ago Changed · -3 lines · -10 tokens per session ad237e54fc70
- 2mo ago Changed · +1 lines b5136bf33615
- 2mo ago First seen · 38 lines · 40 tokens per session scan A ffdc9b7ef9e0
dsh-web-agent-coding is a skill published in the GitHub repository zhu1090093659/dsh-web (10,120 stars, last pushed yesterday), licensed Apache-2.0. It adds 30 tokens to every session and 935 once invoked, about $0.0001 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 2026-08-30.
dsh-web-agent-coding compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| dsh-web-agent-codingzhu1090093659/dsh-web | Skill | 10k | yesterday | Apache-2.0 | A | 30 | Claude Code |
| manage-taskboardshengsheng90/DSH-taskboard | Skill | 326 | 4d ago | Apache-2.0 | A | 88 | Codex |
| industry-research-methodPerryLink/dsh-industry-research | Skill | 216 | yesterday | Apache-2.0 | A | 96 | Claude Code, Codex |
| company-research-methodPerryLink/dsh-industry-research | Skill | 216 | yesterday | Apache-2.0 | A | 83 | Claude Code, Codex |
| dsh-plugin-guidePerryLink/dsh-plugin-guide | Skill | 64 | 2d ago | Apache-2.0 | A | 62 | Claude Code, Codex |
| setup-matt-pocock-skillsgongyijie85/mattpocock-skills-dsh | Skill | 29 | 1mo ago | MIT | A | 44 | Claude Code |
| wayfindergongyijie85/mattpocock-skills-dsh | Skill | 29 | 1mo ago | MIT | A | 46 | Claude Code |