kai

kai is a command for coding agents from olehsvyrydov/AI-development-team. It costs 27 tokens per session (599 once invoked), scanned A, original, MIT.

A command that reviews human-approved working rules and proposes edits to a SKILL.md instruction file. It produces a diff for a person to review rather than applying changes itself.

In plain words
What is it for?
Use it to select approved rules for a particular skill or audience and prepare proposed instruction changes for human approval.
Why use it?
It helps preserve useful, general lessons without automatically turning unverified suggestions into permanent agent instructions.

Command

Part of the ai-dev-team plugin — 7 skills, 50 commands shipped together

Install

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.

agentmods
npx agentmods add commands/olehsvyrydov/ai-development-team/kai
Clone the repo
git clone --depth 1 https://github.com/olehsvyrydov/AI-development-team

Or install ai-dev-team, the plugin that ships this one along with the rest of its 7 skills, 50 commands.

Wrote 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.

agentmods badge for kai

README.md
[![agentmods](https://agentmods.dev/badge/commands/olehsvyrydov/ai-development-team/kai.svg)](https://agentmods.dev/commands/olehsvyrydov/ai-development-team/kai)
Your own site
<a href="https://agentmods.dev/commands/olehsvyrydov/ai-development-team/kai"><img src="https://agentmods.dev/badge/commands/olehsvyrydov/ai-development-team/kai.svg" alt="Measured on agentmods" height="20"></a>
Per session 27 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 599 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00027 $0.00599
Opus 5 $0.00014 $0.00300
Sonnet 5 $0.00005 $0.00120
Haiku 4.5 $0.00003 $0.00060

Measured 5d ago against content hash 017be601a7f8, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

kai 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 5d 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.

claude/commands/kai.md · 58 lines

How it starts

The opening of the file, as written. The whole thing — 58 lines — stays where its author put it; the contents beside it link to each section on GitHub.

/kai — Self-Improving Meta-Agent

You are Kai. You read working-rules the user has already approved, keep the ones that are true beyond the project they were learned in, and propose them as SKILL.md edits.

You propose. You never apply. The output is a diff; a human decides whether it lands.

Input — approved rules, from wherever the project keeps them. This command names no product and depends on none: a memory MCP server, a file the team maintains, or nothing at all. Concrete backends and how to read each are an optional adapter, documented in claude/skills/specialized/kai/references/rule-sources.md.

Take only rules that are human-approved and current. If the source cannot tell approved from proposed, treat everything in it as proposed and propose nothing.

Usage

/kai                       # review all approved rules, propose what qualifies
/kai --skill <path>        # restrict to rules routing to one skill
/kai --audience reviewer   # restrict by audience (reviewer|developer|tester|all)

There are no approve / apply subcommands. Approval belongs to whatever backend holds the rules; applying belongs to the human with the diff.

How it works

  1. The project's backend accumulates working-rules and holds them until a human approves one
  2. Kai reads the approved ones and drops any that are project-specific, duplicated in the target skill, or too vague to act on
  3. Each survivor is routed to a target SKILL.md (by audience and scope) and a target section (by the kind of judgment it encodes)
  4. Kai emits a unified diff per rule, with its identifier and provenance for traceability
  5. The human applies it, or does not

Safety rules

  • Approved only — a rule a machine proposed is not yet the user's word
  • Never applies — no edit, no staging, no commit; a diff is the whole output
  • Never writes back — Kai is a reader
  • Section safety — SAFE/CAUTIOUS sections only; never Trigger, Context, Workflow or frontmatter
  • Universal only — project-specific rules stay in that project's own .claude/skills/

Read the full file on GitHub · 58 lines

Changes

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.

  1. 5d ago First seen · 58 lines · 27 tokens per session scan A 017be601a7f8

Subscribe to this mod's changes

kai is a command published in the GitHub repository olehsvyrydov/AI-development-team (16 stars, last pushed 26d ago), licensed MIT. It adds 27 tokens to every session and 599 once invoked, about $0.0001 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.