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 instructions/aibenyclaude-coder/kira/claude-mdgit clone --depth 1 https://github.com/aibenyclaude-coder/KiraWrote 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/instructions/aibenyclaude-coder/kira/claude-md)<a href="https://agentmods.dev/instructions/aibenyclaude-coder/kira/claude-md"><img src="https://agentmods.dev/badge/instructions/aibenyclaude-coder/kira/claude-md.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 | $0.00639 | $0.00639 |
| Opus 5 | $0.00319 | $0.00319 |
| Sonnet 5 | $0.00128 | $0.00128 |
| Haiku 4.5 | $0.00064 | $0.00064 |
Grade A, and why
Kira CLAUDE.md 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.
How it starts
The opening of the file, as written. The whole thing — 80 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Kira — Project Rules
Where agents shine.
What is Kira
MCP server that auto-manages Skills (how to do it) and Scars (what to avoid) for AI agents. One install, zero maintenance.
Architecture
src/— TypeScript MCP serverskills/community/— Community skills (JSON)skills/vendor/— Vendor skills (JSON, 審査制)skills/scars/— Scars: past failure patterns (JSON)DESIGN.md— Design philosophy and decisionsPLAN.md— Full business plan
Skill JSON Schema
{
"id": "community.<slug>.v1",
"keywords": ["firing keyword"],
"contexts": ["nextjs", "react"],
"title": "Human-readable title",
"summary": "One-line.",
"source": "community",
"declaration": "What agent announces before executing.",
"instructions": "## Markdown step-by-step",
"version": "1.0.0",
"updated_at": "2026-04-11T00:00:00Z"
}
Scar JSON Schema
{
"id": "scar.<slug>.v1",
"keywords": ["firing keyword"],
"contexts": ["nextjs"],
"title": "What went wrong",
"summary": "One-line.",
"severity": "critical",
"mistake": "What the agent did wrong.",
"instead": "What to do instead.",
"hit_count": 100,
"version": "1.0.0",
"updated_at": "2026-04-11T00:00:00Z"
}
Development Rules
- Build check: Always run
npm run buildafter editing src/ - Demo check: Always run
npm run demoafter any change to verify all skills load - Skill quality: Every skill must have Common Errors + What NOT to do sections
- Scar quality: Every scar must have concrete mistake + concrete instead
- Keywords: Include at least 3 keyword variants per skill
- Contexts: Always specify project context tags
- No code in skills: Skills are natural language Markdown instructions only
- Validate JSON: All JSON files must be valid (run
python3 -c "import json; json.load(open('path'))")
Commit Convention
- Skill additions:
add skill: <title> - Scar additions:
add scar: <title> - Server changes:
server: <description> - Infrastructure:
infra: <description>
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.
- 5d ago First seen · 80 lines · 639 tokens per session scan A 2069c3704020
Kira CLAUDE.md is an instructions file published in the GitHub repository aibenyclaude-coder/Kira (0 stars, last pushed 1mo ago), licensed MIT. It adds 639 tokens to every session, about $0.0032 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 instructions, from other repositories
mcpm.sh GEMINI.md
Instructions for pathintegral-institute/mcpm.sh: Reference CLAUDE.md for project conventions.
agents-in-a-box AGENTS.md
AGENTS.md instructions for stevengonsalvez/agents-in-a-box, covering agent instructions, communication, beads, quick reference and landing the plane (session completion).
sample-well-architected-skills-and-steering copilot-instructions.md
Copilot instructions for aws-samples/sample-well-architected-skills-and-steering, covering well-architected framework, role, when to apply, pillars and design principles.
sessionmem AGENTS.md
AGENTS.md instructions for catfish-1234/sessionmem, covering sessionmem: persistent memory, startup, when to store memories (storememory), when to retrieve memories mid-session (retrievememories) and at session end (recommended: do this without being asked).
margaret AGENTS.md
Instructions for sofiahernandes/margaret: You're a principal engineer who sizes builds to the actual requirement, not the one someone might need later. Clean falls out of good design; it isn't a separate pass.
syncai copilot-instructions.md
Instructions for dmtrkzntsv/syncai, covering claude.md, build & run, architecture, lifecycle and identify is the routing decision.