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/knuckles-team/repository-manager/agents-mdgit clone --depth 1 https://github.com/Knuckles-Team/repository-managerWrote 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/knuckles-team/repository-manager/agents-md)<a href="https://agentmods.dev/instructions/knuckles-team/repository-manager/agents-md"><img src="https://agentmods.dev/badge/instructions/knuckles-team/repository-manager/agents-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 | $0.18262 | $0.18262 |
| Opus 5 | $0.09131 | $0.09131 |
| Sonnet 5 | $0.03652 | $0.03652 |
| Haiku 4.5 | $0.01826 | $0.01826 |
Grade A, and why
repository-manager AGENTS.md scanned grade A with 1 finding 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 4d 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.
Runs shell commandslowCapability
Expected in a hook, worth knowing in a rule or an instructions file.
`subprocess.run`, so a declared command cannot reach the shell without the How it starts
The opening of the file, as written. The whole thing — 1,217 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Repository Manager Agent Documentation
Claude Code loads this file via
CLAUDE.md(@AGENTS.mdimport) — the two stay in sync. Edit this file, notCLAUDE.md.
This document provides an overview of the Repository Manager agent, its architecture, and how to use it.
Tech Stack & Architecture
- Language: Python 3.11–3.14
- Core Framework: Pydantic AI & Pydantic Graph
- Tooling:
requests,pydantic,pyyaml,python-dotenv,fastapi,llama_index,FastMCP - Architecture: Centered around the
create_agentfactory fromagent-utilities, which has been modernized to support a Unified Skill Loading model (skill_types) and automated Graph Orchestration. - Specialist Discovery: Automated discovery of domain specialist agents from
NODE_AGENTS.md(local) andA2A_AGENTS.md(remote) registries, enabling dynamic graph expansion without hardcoded nodes. - Key Principles:
- Functional and modular utility design.
- Standardized workspace management (
IDENTITY.md,MEMORY.md). - Elicitation First: Robust support for structured user input during tool calls, bridging MCP and Web UIs.
Package Relationships
The Repository Manager agent is built on top of the agent-utilities package, which provides the core Python engine for LLM orchestration, tool execution, and the SSE streaming protocol.
- Backend (
agent-utilities): Handles LLM orchestration, tool execution, and the SSE streaming protocol. - Web Frontend (
agent-webui): A React application that provides a cinematic chat interface and specialized UI components. - Communication: Frontends talk to Backend via SSE for output and standard REST (POST) for input and elicitation responses.
Validation & Diagnostics
To ensure the Repository Manager specialist and its graph lifecycle are functioning correctly, use the following validation tools:
End-to-End Specialist Validation
High-fidelity testing of individual specialist nodes through the SSE streaming protocol. This bypasses the Web UI and provides granular execution logs to monitor tool calls and result registration.
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.
- 4d ago First seen · 1,217 lines · 18,262 tokens per session scan A abba4e953146
repository-manager AGENTS.md is an instructions file published in the GitHub repository Knuckles-Team/repository-manager (2 stars, last pushed 6d ago), licensed MIT. It adds 18,262 tokens to every session, about $0.0913 per session on Opus 5. A static security scan graded it A with 1 finding (runs shell commands). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other instructions, from other repositories
blockrun-mcp AGENTS.md
AGENTS.md instructions for BlockRunAI/blockrun-mcp, covering blockrun mcp, commands, project structure, key dependencies and install in codex.
openrouter-mcp-multimodal AGENTS.md
AGENTS.md instructions for stabgan/openrouter-mcp-multimodal, covering agent instructions, before you ship, releasing (read this before publishing), short version and version files (must all match package.json).
intervals-icu-mcp CLAUDE.md
Instructions for hhopke/intervals-icu-mcp, covering claude.md, project overview, development commands, architecture (quick reference) and tool categories.
ai-toolkit AGENTS.md
AGENTS.md instructions for pipefy/ai-toolkit, covering repository guidelines, documentation map, project structure, import namespace migration: pipefysdk → pipefy and src/pipefysdk/init.py (transitional shim).
flyto-core CLAUDE.md
Claude Code instructions for flytohub/flyto-core, covering claude notes, cross-agent handoff and shared code intelligence.
Plonk AGENTS.md
AGENTS.md instructions for ostapondo/Plonk, covering agent rules, layout, adding a module, build & verify and code style.