instruction-loading

An instruction-loading system that combines runtime rules, an agent’s instruction files, and the current run’s context before the agent starts work.

In plain words
What is it for?
Preparing prompts for agent runs, loading skills and wake-up context, and recording which instruction layers were used or omitted.
Why use it?
It helps agents resume with the right identity, boundaries, workspace, resources, and task details while leaving a reviewable record of what they saw.

Agent

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 agents/undertone0809/rudder/instruction-loading
Clone the repo
git clone --depth 1 https://github.com/Undertone0809/rudder
Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 9,407 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 $0.00000 $0.09407
Opus 5 $0.00000 $0.04704
Sonnet 5 $0.00000 $0.01881
Haiku 4.5 $0.00000 $0.00941

Measured 2d ago against content hash 94d14444c1ae, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

instruction-loading 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 2d 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.

doc/product/domains/agents/instruction-loading.md · 715 lines

How it starts

The opening of the file, as written. The whole thing — 715 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Agent Instruction Loading

AGENT.INSTRUCTIONS.001

Contract Summary

Rudder must assemble each runtime agent's instruction frame from three kinds of truth before invoking the provider:

  • durable runtime-owned operating rules
  • durable agent-owned instruction files
  • dynamic run context for the current scene, workspace, resources, skills, and wake reason

The contract protects both the content and the ordering of that frame. A run handoff must let a reviewer answer: what did the agent see, what was intentionally omitted, what runtime context was persisted, and what evidence shows the instruction stack used for this run.

Intent / User Job

Operators rely on Rudder agents to resume work without rediscovering their identity, boundaries, workspace, resources, issue context, and runtime-specific rules on every run. Reviewers rely on the run record to explain why an agent acted as it did.

This contract exists so future development can safely change agent runtimes, skills, wakeups, resources, or prompt templates without accidentally removing a loaded instruction layer, duplicating dynamic context, or adding heartbeat-only behavior to comment-triggered issue work.

Why / Design Reasoning

Instruction loading is split by ownership:

  • Rudder runtime code owns the universal operating contract and heartbeat-only runtime instruction. These rules are stable platform behavior and cannot depend on mutable agent home files.
  • The agent's instruction directory owns durable role/persona/tool/memory material. These files let an operator customize an agent without editing runtime code.
  • The run context owns per-run workspace, resources, startup context, session handoff, issue/comment context, and wake reason. These sections must be late enough to be current, but early enough for prompt templates and providers to consume once.

The key tradeoff is explicit ordering over template convenience. Rudder moves resource context into the shared instruction prefix, then clears duplicate template aliases so adapters do not re-inject the same resource block later. Heartbeat instructions are runtime-owned instead of read from legacy HEARTBEAT.md because heartbeat behavior is a Rudder contract, not an agent-local note. They are loaded only for rudderScene=heartbeat. Issue, review, chat, and automation runs are excluded so task assignment, review, comment mention, chat, and automation work are not framed as a generic autonomous heartbeat loop.

Read the full file on GitHub · 715 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. 2d ago First seen · 715 lines · 0 tokens per session scan A 94d14444c1ae

Subscribe to this mod's changes

instruction-loading is an agent published in the GitHub repository Undertone0809/rudder (287 stars, last pushed 2d ago), licensed Apache-2.0. It costs nothing until one of its globs matches a file; then it loads 9,407 tokens. 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.

Related

Other agents, from other repositories

Demonstrate

Agent for demonstrating VS Code features.

microsoft/vscode · 10 tokens

playwright-test-generator

Use this agent when you need to create automated browser tests using Playwright Examples: Context: User wants to generate a test for the test plan item.

microsoft/playwright · 151 tokens

.NET-Notebook-Migration-Agent

Expert .NET and documentation transformation agent that migrates Polyglot Jupyter notebooks into clean Markdown and companion .NET sample code.

microsoft/ai-agents-for-beginners · 33 tokens

AVM Owner Triage

Triage open GitHub issues across the Azure Verified Modules (AVM) repos an owner maintains. Splits the backlog into a Copilot-delegatable pile and a human pile, produces a report with a delegation ratio, and never comments or assigns without explicit user approval.

github/awesome-copilot · 61 tokens

Ultimate Transparent Thinking Beast Mode

Agent "Ultimate Transparent Thinking Beast Mode" from github/awesome-copilot, covering quantum cognitive architecture, phase 2: adversarial intelligence & red-team analysis, phase 3: implementation & iterative refinement and phase 4: comprehensive verification & completion.

github/awesome-copilot · 11 tokens

code-reviewer

Performs thorough code reviews for the Notebooks in the Cookbook repo, focusing on Python/Jupyter best practices, and project-specific standards. Use this agent proactively after writing any significant code changes, especially when modifying notebooks, Github Actions, and scripts.

anthropics/claude-cookbooks · 52 tokens