Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/jayvee/aigonnpx agentmods add skills/jayvee/aigon/code-tourWrote 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/jayvee/aigon/code-tour)<a href="https://agentmods.dev/skills/jayvee/aigon/code-tour"><img src="https://agentmods.dev/badge/skills/jayvee/aigon/code-tour.svg" alt="Measured on agentmods" 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 | $0.00074 | $0.01754 |
| Opus 5 | $0.00037 | $0.00877 |
| Sonnet 5 | $0.00015 | $0.00351 |
| Haiku 4.5 | $0.00007 | $0.00175 |
Grade A, and why
code-tour 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 — 187 lines — stays where its author put it; the contents beside it link to each section on GitHub.
code-tour
docs/code-tour.md is the annotated guided reading of Aigon's core logic — 25 core
sections built around verbatim source excerpts, in the order the machinery runs, each
with commentary on what it does and why it is shaped that way. Its reader is someone who
does not routinely read the code and wants an accurate mental model of the system.
You are its maintainer. It is part of the internal doc set, alongside docs/architecture.md.
This file is plain markdown and agent-agnostic. Non-Claude agents: read and follow it directly; there is nothing Claude-specific below.
When to act
Update the tour when:
- You changed a file the tour quotes — even a whitespace edit above the excerpt shifts the line anchor.
- You changed the behaviour of a quoted function, whether or not the lines moved.
- You added or substantially reworked a subsystem that a new reader would need to understand the system (see Adding a section — the bar is high).
- You removed or renamed something the tour references.
Do not touch it when:
- Your change is in a file the tour doesn't quote and doesn't alter any pattern it teaches.
- You "improved the prose". The commentary is deliberate; don't churn it.
- You want to document a bugfix. That belongs in the spec's implementation log, not here.
Check first, always:
node scripts/check-code-tour.js
Validates that every JavaScript excerpt has a dedicated, aligned file:line anchor and
that every retained excerpt segment is byte-identical to its source. It also checks all
inline anchors for missing files and out-of-range lines. Run this before you decide
whether there is work to do — often there isn't.
The rules
These are what make the document trustworthy. Breaking them silently is worse than leaving the tour stale, because a reader who can't trust one excerpt can't trust any of them.
1. Excerpts are verbatim. Copy the real bytes out of the real file. Never retype from
memory, never tidy, never rename a variable to read better, never fix a typo in a comment.
If the code is ugly, the tour shows ugly code. Elision is the only permitted edit, and
only as a standalone // … line where you dropped something.
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 · 187 lines · 74 tokens per session scan A 80c591c1f4be
code-tour is a skill published in the GitHub repository jayvee/aigon (25 stars, last pushed 5d ago), licensed Apache-2.0. It adds 74 tokens to every session and 1,754 once invoked, about $0.0004 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
printing-press-retro
Use when the user asks to retro, run a retrospective, file findings, or improve the Printing Press after a printed-CLI run. Trigger phrases: "retro", "retrospective", "what went wrong", "improve the press", "post-mortem", "lessons learned", "what can we improve", "file a retro", "submit findings". Do not use for…
pronounce-word
User asks how to pronounce an English word or project/product name ("how to pronounce X", "pronounce X", "X 怎么读", "X 怎么发音", "读一下 X"). Generate audio via the say-it CLI so the user actually HEARS the word — three times by default — instead of only writing IPA/syllable hints. The CLI consults a community-maintained…
教程类H5-小红书小工具
用 AI 生图 + SPA 课程架构 + 离线 H5 zip = 在小红书"创作者中心 → 小工具"发布的多课程 H5 互动产品。讲清楚 4 层栈 / 5 步流程 / 7 节课程都适用。触发词：「做教程类 H5」「小红书小工具」「H5 课程」「离线 H5 课程」「多节课 H5」。.
sansheng-distill
Distill any book/text into a self-contained HTML page with 5-layer structure (overview / per-chapter / book-soul / action-checklist / credibility-grading). Zero external dependencies, embedded mermaid mind-map, thinking-process traceable. Use when user says "拆书", "读书蒸馏", "AI 拆书", "把一本书蒸透", "把 [book] 拆成 5 层", or any…
dos-goal-fleet
Launch multiple goal-scoped workers in waves, each with a witness-gated stop condition and dos arbitrate file-tree safety. Use when an operator asks to run several independent goals in parallel and fold only verified ships.
dos-goal-gate
Ground a keep-working goal in evidence the worker did not author by wiring dos hook stop to refuse false done claims. Use for one self-stopping agent or loop worker; use dos-witness-claim for fold barriers.