Wanwu is an enterprise platform for building AI agents, workflows, retrieval-augmented applications, and managing models in multi-tenant environments. It is designed for developers and enterprise teams delivering AI applications and integrations. The catalogue entries provide skills and agents for using the platform.
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 UnicomAI/wanwu --skill create-bkngit clone --depth 1 https://github.com/UnicomAI/wanwuWrote 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/unicomai/wanwu/create-bkn)<a href="https://agentmods.dev/skills/unicomai/wanwu/create-bkn"><img src="https://agentmods.dev/badge/skills/unicomai/wanwu/create-bkn/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/unicomai/wanwu/create-bkn"><img src="https://agentmods.dev/badge/skills/unicomai/wanwu/create-bkn.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 1 finding, up to high
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- high System Prompt Leakage · line 118 Skill contains instructions that could directly expose system prompts, internal rules, or hidden instructions to users or external parties.Fix: Remove any instructions that reveal, print, or output system prompts or internal rules. System instructions should never be exposed to end users.
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 | $0.02432 |
| Opus 5 | $0.00049 | $0.01216 |
| Sonnet 5 | $0.00020 | $0.00486 |
| Haiku 4.5 | $0.00010 | $0.00243 |
Grade A, and why
create-bkn 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 6d 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 — 132 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Create BKN
Generate well-formed BKN directories (Markdown + YAML frontmatter) per v2.0.1.
Works with ontology-core
create-bkn authors the .bkn tree; ontology-core runs ontology bkn push / pull after files exist.
What is BKN
BKN is Markdown + YAML frontmatter for schema; one file per definition under typed subfolders. Details (sections, required tables, types) live in references/SPECIFICATION.llm.md.
Directory layout
{network_dir}/
├── SKILL.md
├── network.bkn
├── CHECKSUM # optional; SDK may generate
├── object_types/
├── relation_types/
├── action_types/
├── concept_groups/
└── data/ # optional CSV instance data
Workflow
- Gather requirements — objects, relations, actions, optional concept groups
- Read spec — references/SPECIFICATION.llm.md (format rules, sections, frontmatter types)
- Pick templates — copy/adapt from assets/templates/ (
network_type.bkn.template,object_type.bkn.template, …) - Create
network.bkn— root file; align with Network Overview- MUST: generate a fresh UUID v4 locally (e.g. Python
uuid.uuid4()) and write it as theidfield in frontmatter at file creation time. Never leaveidempty,null,~, or absent —ontology bkn validate/pushboth require a non-empty string id, and the bkn-creator flow does not callontology bkn createto acquire a server-assigned id. - The locally generated UUID is the final
kn_id; any other.bknfile that references the network id (e.g.network_idinobject_types/*.bkn) must reuse the same UUID.
- MUST: generate a fresh UUID v4 locally (e.g. Python
- Create
object_types/*.bkn— one file per object,{id}.bkn - Create
relation_types/*.bkn— one file per relation - Create
action_types/*.bkn— optional, may be empty. One file per genuine write-operation on instances. Decide per the rules below; if none apply, create no files and leaveaction_types/and the Network Overview ActionType list empty (keep them consistent).- Build an action when the PRD contains an explicit write-operation (create/update/delete of instances) with an identifiable trigger condition or a bound tool.
- Create none when the PRD is read-only (query / monitor / analysis / stats / trace) — those belong to object-type query / subgraph / metric / semantic search, not ActionType.
- Ambiguous cases (e.g. whether "approval" is a write-action vs. a status transition) — surface to the user for confirmation; never silently build or silently skip.
- Create
concept_groups/*.bkn— optional - Update
network.bkn— list all IDs in Network Overview - Add root
SKILL.mdin the BKN directory — same folder asnetwork.bkn(this is not the create-bkn skill file); agent-facing guide for that network (see Delivered BKN: root SKILL.md) - Review (MUST) — cross-check Validation checklist and Business rules placement; fix IDs, cross-refs, headings
- 特别核对
action_types/*.bkn每个文件 frontmatter 是否含action_type: add|modify|delete这一行(位置在 frontmatter,不在 markdown body 的 Bound Object 表),遇到查询/监控/追溯/校验等只读语义,删掉对应 ActionType 文件并在network.bkn的 Network Overview 同步移除该 id
- 特别核对
- Validate (MUST) —
ontology bkn validate <dir>(see Validation) - Import (optional) —
ontology bkn push <dir>
What ships with it
17 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.
- assets/templates/action_type.bkn.template 1.7 KB
- assets/templates/concept_group.bkn.template 266 B
- assets/templates/network_type.bkn.template 442 B
- assets/templates/object_type.bkn.template 1.3 KB
- assets/templates/relation_type.bkn.template 841 B
- references/examples/k8s-network/action_types/cordon_node.bkn 1.4 KB
- references/examples/k8s-network/action_types/restart_pod.bkn 1.4 KB
- references/examples/k8s-network/CHECKSUM 487 B
- references/examples/k8s-network/concept_groups/k8s.bkn 495 B
- references/examples/k8s-network/GUIDE.md 2.9 KB
- references/examples/k8s-network/network.bkn 513 B
- references/examples/k8s-network/object_types/node.bkn 1.3 KB
- references/examples/k8s-network/object_types/pod.bkn 1.1 KB
- references/examples/k8s-network/object_types/service.bkn 1.2 KB
- references/examples/k8s-network/relation_types/pod_belongs_node.bkn 415 B
- references/examples/k8s-network/relation_types/service_routes_pod.bkn 430 B
- references/SPECIFICATION.llm.md 9.6 KB
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.
- 6d ago First seen · 132 lines · 98 tokens per session scan A 91cbadecb218
create-bkn is a skill published in the GitHub repository UnicomAI/wanwu (2,461 stars, last pushed 5d ago), licensed Apache-2.0. It adds 98 tokens to every session and 2,432 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 2026-09-03.
Other skills, from other repositories
dify
Use when building LLM applications with visual workflow — RAG knowledge bases, AI agents, chatbots with drag-and-drop orchestration. Dify: open-source LLM app platform supporting 30+ models (OpenAI, Claude, DeepSeek, Ollama, Qwen, GLM) with Docker deployment.
ai-skills
Use when building LLM applications, RAG knowledge bases, AI agents, terminal coding agents, multi-model orchestration, plugin-based agent harnesses, or file translation. Index of 10 skills: Dify, Hermes Agent, OpenClaw, OpenCode, Pi, DocuTranslate, Oh-My-OpenAgent, Superpowers-zh, DeepSeek Harness, My OpenCode…
Web2Skill
Convert one public website URL or an explicit batch of public URLs into a reusable skill zip backed by rendered HTML snapshots and a bounded JSONL retrieval index. Use to discover a documentation directory from one URL, crawl a supplied URL set sequentially, generate a source profile, or package indexed web content as…
Book2Skill
Convert one or more TXT, Markdown, DOCX, or PDF documents into a reusable skill zip backed by normalized Markdown, extracted images, and a grounded JSONL knowledge index. Invoke this skill before inspecting task files, then execute its workflow directly without listing directories.
agent-orchestration-multi-agent-optimize
Optimize multi-agent systems with coordinated profiling, workload distribution, and cost-aware orchestration. Use when improving agent performance, throughput, or reliability.
ai-agents-architect
Expert in designing and building autonomous AI agents. Masters tool use, memory systems, planning strategies, and multi-agent orchestration.