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 agentmods add skills/cogni-work/insight-wave/knowledge-querynpx skills add cogni-work/insight-wave --skill knowledge-querygit clone --depth 1 https://github.com/cogni-work/insight-waveWhat 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 | $0.00124 | $0.05164 |
| Opus 5 | $0.00062 | $0.02582 |
| Sonnet 5 | $0.00025 | $0.01033 |
| Haiku 4.5 | $0.00012 | $0.00516 |
Grade A, and why
knowledge-query 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 3d 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 — 353 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Knowledge Query
Ask a question against a bound cogni-knowledge base. This skill is the shallow
rung of the query↔research depth ladder: it ranks the pages that cover the
question, reads them, and synthesizes a grounded, cited answer — one question,
index-first, ≤12 pages, no web, no verify (the deep wiki-only report rung is
owned by knowledge-compose --source wiki).
It runs natively on the vendored wiki engine — it consumes the shared
wiki-grounding.py discovery primitive directly and reads the wiki pages with
the Read tool, so a Karpathy base answers questions without cogni-wiki
installed. The cogni-knowledge value-add:
- Binding-aware wiki path resolution — no
--wiki-rootfrom the user; read frombinding.json. - Index-first grounded synthesis — rank covering pages via
wiki-grounding.py, read them, answer only from what they say (with[[slug]]citations), and report honestly when coverage is thin. - Knowledge-base footer — every answer ends with one line tying it to the knowledge slug + deposit count + fetch-cache health, so the user remembers where the answer came from and how much evidence the base holds.
Read ${CLAUDE_PLUGIN_ROOT}/references/differentiation-thesis.md once per session to anchor on the accumulation thesis. The synthesis discipline below — answer only from the pages you read, cite every claim, admit uncertainty — is the shallow rung's core contract.
When to run
- User asks a question and references a knowledge base by slug — "what does my eu-ai-act base know about X?", "query the eu-ai-act base on Article 6"
- User explicitly invokes
/cogni-knowledge:knowledge-query - User asks a question that clearly lives in a bound knowledge base's domain after
knowledge-resumehas shown what is in the base
Never run when
- No
binding.jsonexists at the resolved knowledge root — route the user to/cogni-knowledge:knowledge-setupfirst. - The user wants a deep, multi-section, web-verified report rather than a single grounded answer — that is
knowledge-compose --source wiki(the deep rung), not this skill.
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.
- 3d ago First seen · 353 lines · 124 tokens per session scan A a3460677f064
knowledge-query is a skill published in the GitHub repository cogni-work/insight-wave (12 stars, last pushed 3d ago), licensed Apache-2.0. It adds 124 tokens to every session and 5,164 once invoked, about $0.0006 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.
Other skills, from other repositories
9router-embeddings
Generate vector embeddings via 9Router /v1/embeddings using OpenAI / Gemini / Mistral / Voyage / Nvidia / GitHub embedding models for RAG, semantic search, similarity. Use when the user wants embeddings, vectors, RAG, semantic search, or to embed text.
azure-search-documents-dotnet
Azure AI Search SDK for .NET (Azure.Search.Documents). Use for building search applications with full-text, vector, semantic, and hybrid search. Covers SearchClient (queries, document CRUD), SearchIndexClient (index management), and SearchIndexerClient (indexers, skillsets). Triggers: "Azure Search .NET"…
embedding-strategies
Select and optimize embedding models for semantic search and RAG applications. Use when choosing embedding models, implementing chunking strategies, or optimizing embedding quality for specific domains.
gpt-image-2
面向 GPT Image 2 的图像生成 / 编辑技能。可在 3 种环境下使用：(A) Garden 本地模式，通过 OpenAI 兼容接口直接出图并落盘；(B) Host-Native 模式，把本 Skill 当作提示词工程指引，把渲染好的 prompt 交给宿主 Agent 自带的图像工具出图；(C) Advisor 模式，宿主无任何图像工具时退化为高质量 prompt 顾问。涵盖 18 大类、80+ 个结构化模板，覆盖海报 / UI / 产品 / 信息图 / 学术图 / 技术架构图 / 漫画 / 头像 / 流程板 / 电影分镜 / IP 周边 / 编辑工作流等场景。.
wax
Swift framework guidance for Wax on-device memory/RAG. Use when writing Swift code with the public Memory facade, experimental PhotoMemory / VideoMemory, BuiltInMultimodalEmbeddings, embedding providers, retrieval modes, or hybrid search. For agent operators using the Wax MCP server tools, use the separate wax-mcp…
session-rag-eval
Run and debug Chatbox session attachment RAG model evaluation with synthetic and real long-file fixtures.