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/ker102/harneloop/agents-mdgit clone --depth 1 https://github.com/Ker102/HarneloopWhat 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 | $0.01014 | $0.01014 |
| Opus 5 | $0.00507 | $0.00507 |
| Sonnet 5 | $0.00203 | $0.00203 |
| Haiku 4.5 | $0.00101 | $0.00101 |
Grade A, and why
Harneloop AGENTS.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 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.
How it starts
The opening of the file, as written. The whole thing — 65 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Agent Instructions For Harneloop
Harneloop is a framework for agent-built, artifact-aware harness units.
Current Stage
This repository is in the first core prototype stage. Keep changes small, tested, and generic. Do not let Blender-specific assumptions leak into the framework core.
First Onboarding Step
When starting a new harness unit, run:
harneloop onboard
If Harneloop is not installed yet, use:
uv run harneloop onboard
Use the five onboarding questions to collect only the minimum context needed for the first baseline attempt: harness goal, usage context, success strategy, validation preference, and environment status. Treat success criteria and artifact choices as guided options. The user does not need to know validation design up front.
Harneloop records environment mappings; it does not magically discover test endpoints, MCP tools, run commands, screenshot locations, render outputs, or artifact paths. The onboarding agent must inspect the actual project/environment, determine how artifacts are produced, and write that mapping into the harness via environment connect, attempt plan, run records, and artifact records.
Every new harness unit includes operational-map.md. Treat it as the current orientation for this specific unit: what it is trying to improve, which systems and tools it touches, what artifacts or evidence are useful, how the environment can usually be run or reset, fragile spots, assumptions, open questions, and where prior evidence lives. It is context and navigation, not a rigid procedure. Update it when the workflow, evidence needs, artifact paths, environment assumptions, or automation strategy change.
Track capability gaps in operational-map.md. Operating-agent capabilities are what the current agent can actually use while building the unit, such as terminal, filesystem, browser, MCPs, package managers, visual inspection, database access, or custom tools. Unit/target-agent tools are the tools designed into the harness unit or provided to the target agent. Keep those layers separate.
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.
- 2d ago First seen · 65 lines · 1,014 tokens per session scan A c9c3ae252aa8
Harneloop AGENTS.md is an instructions file published in the GitHub repository Ker102/Harneloop (4 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 1,014 tokens to every session, about $0.0051 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
best-of-Agent-Harnesses CLAUDE.md
Claude Code instructions for RyanAlberts/best-of-Agent-Harnesses, covering repo notes for ai coding agents, commit identity (required), branching and pushing (required), verifying contribution credit (required) and how the list is generated.
keen-agent AGENTS.md
Instructions for mochow13/keen-agent, covering keen agent, important guidelines, architecture, permission system and cast to gif conversion.
keen-agent CLAUDE.md
Instructions for mochow13/keen-agent, a project described as: Terminal agent harness built from a coding-agent foundation — plug in your own agent via YAML config.
eigent AGENTS.md
AGENTS.md instructions for eigent-ai/eigent, covering eigent agent instructions, ui work, product terminology and validation.
HARTOS CLAUDE.md
Instructions for hertz-ai/HARTOS, covering claude.md, project overview, master key - ai exclusion zone, branch discipline — main branch only (mandatory) and sibling repos — canonical filesystem paths.
openacme CLAUDE.md
Instructions for sandydasari/openacme, covering openacme — claude code guide, workspace layout, the agent loop — request path, why sse-only and custom data parts.