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/hogu-dev/llmkb/kb-querynpx skills add hogu-dev/llmkb --skill kb-querygit clone --depth 1 https://github.com/hogu-dev/llmkbWhat 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 | $0.00050 | $0.00384 |
| Opus 5 | $0.00025 | $0.00192 |
| Sonnet 5 | $0.00010 | $0.00077 |
| Haiku 4.5 | $0.00005 | $0.00038 |
Grade A, and why
kb-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 yesterday.
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.
What it actually says
kb-query
Answer a question from one or more llmkb knowledge bases, cite what you used,
and — when the answer produced durable new knowledge — file it back.
When to use
A user asks a question that the KB(s) should be able to answer, or asks you to research within them.
Steps
- Route via the index first. Read
wiki/index.md(each KB's) — one line per page — to find the pages most likely to hold the answer. For multiple KBs, read each index; results are tagged by KB name. - Search at scale. If the index is large or routing is unclear, run
llmkb search "<query>" --kb <dir>(repeat--kbfor several KBs). Use the hits to pick pages. - Read the pages you selected and compose the answer.
- Cite every page you drew on by its
[[Title]]. - Compound (optional but encouraged). If answering produced durable new
synthesis not already captured, write it as a
conceptpage (emptysources= agent-authored), then runllmkb reindex, append a## [YYYY-MM-DD] synthesis | <title>line towiki/log.md, and runllmkb lint.
Done when
The question is answered with page citations, and any new synthesis is filed, indexed, logged, and lint-clean.
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.
- yesterday First seen · 36 lines · 50 tokens per session scan A 714c8f1c0e02
kb-query is a skill published in the GitHub repository hogu-dev/llmkb (1 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 50 tokens to every session and 384 once invoked, about $0.0003 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
recursive-decomposition
Handle tasks that exceed the context window by decomposing them: size and filter the input, chunk it, run recursive sub-agents on independent parts, verify on small windows, and synthesise programmatically, following the Recursive Language Models (RLM) research by Zhang, Kraska and Khattab (2025). Use when a task…
agentica-sdk
Build Python agents with Agentica SDK - @agentic decorator, spawn(), persistence, MCP integration.
agentica-claude-proxy
Guide for integrating Agentica SDK with Claude Code CLI proxy.
ast-grep-find
AST-based code search and refactoring via ast-grep MCP.
braintrust-analyze
Analyze Claude Code sessions via Braintrust.
agent-orchestration
When the user asks to implement something, use implementation agents to preserve main context.