Borrowing it
Nothing to install: this file belongs to guoxiaoshuai2023/oh-my-harness. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/guoxiaoshuai2023/oh-my-harness/main/AGENTS.mdgit clone --depth 1 https://github.com/guoxiaoshuai2023/oh-my-harnessWrote 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/guoxiaoshuai2023/oh-my-harness/agents-md)<a href="https://agentmods.dev/instructions/guoxiaoshuai2023/oh-my-harness/agents-md"><img src="https://agentmods.dev/badge/instructions/guoxiaoshuai2023/oh-my-harness/agents-md/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/instructions/guoxiaoshuai2023/oh-my-harness/agents-md"><img src="https://agentmods.dev/badge/instructions/guoxiaoshuai2023/oh-my-harness/agents-md.svg" alt="Reviewed on agentmods" width="80" 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 | $0.01298 | $0.01298 |
| Opus 5.5 | $0.00519 | $0.00519 |
| Sonnet 5.5 | $0.00260 | $0.00260 |
| Haiku 4.5 | $0.00130 | $0.00130 |
Grade A, and why
oh-my-harness 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 2mo 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 — 62 lines — stays where its author put it; the contents beside it link to each section on GitHub.
oh-my-harness Repo Router
This is the repository router for oh-my-harness. The default is one adaptive, main-thread-controlled Harness: task facts select the smallest useful subset of nine optional capabilities while every triggered gate remains owned and evidenced. LOW semantic-risk mechanical work stays direct and lightweight.
Always-On Hard Gates
- Do not expand scope, redesign, refactor, format, clean up, or repair adjacent issues unless the accepted task explicitly requires it.
- Do not read, output, save, or modify secrets, credentials, tokens, cookies, localStorage, payment, billing, permission, keys, private environment values, or other sensitive material.
- Do not perform external writes, publish, destructive change, migration, permission/payment/key action, bulk write, or irreversible action unless the exact target/effect is scoped, a plan or dry-run precedes it, and required confirmation is explicit.
- Do not add dependencies unless necessary and explicitly approved.
- Preserve user work. Baseline a dirty worktree and distinguish pre-existing changes before claiming scope compliance.
- Every changed line must trace to the request, accepted Task/Design, main-issued Boundary, or a necessary consequence.
- Never fabricate evidence, validation, review, readback, external action, or completion.
- Never claim completion without proportional validation. Passing tests are evidence, not proof of all real-world behavior.
Routing Table
Read a routed document only when its factual trigger applies. Traceability artifacts are evidence, not route-table entries.
| Trigger | Route |
|---|---|
| Routing-doc index, trigger convention, route maintenance, or route/evidence distinction | docs/agent-routing/README.md |
| Harness use, capability selection, topology revision, invocation, Boundary, dependency, retry, cancellation, synthesis, or independent verification | task-docs/_harness/adaptive-orchestration-protocol.md |
| External system, production/SaaS/database, credential, secret, auth/token/cookie/localStorage, billing/payment, permission, or key boundary | docs/agent-routing/external-systems-and-secrets.md |
| Existing user-visible surface, generated artifact, live/current state, freshness, or stale-evidence risk | docs/agent-routing/current-state-evidence.md |
| Publish, deploy, delete, migration, restore, bulk write, irreversible, permission, billing/payment, key/secret, or high-impact action | docs/agent-routing/high-risk-actions.md |
| Planning/revising task-docs, task scale, delivery-unit boundary, gate-versus-task, or review-only-task risk | docs/agent-routing/task-planning-scale.md |
| Code, documentation, template, generated asset, implementation boundary, minimal diff, or dependency decision | docs/agent-routing/implementation-boundary.md |
| Validation, evidence, implementation report, test strategy, manual verification, maintenance entry, or final reporting | docs/agent-routing/validation-and-reporting.md |
| MEDIUM/HIGH Semantic Risk, cross-surface work, subjective quality, semantic drift, Pass A/Pass B, Design review, Result QA, or AC-pass-but-user-fail risk | task-docs/_harness/semantic-fidelity-protocol.md |
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.
- 2mo ago First seen · 62 lines · 1,298 tokens per session scan A edeef9f2687f
oh-my-harness AGENTS.md is an instructions file published in the GitHub repository guoxiaoshuai2023/oh-my-harness (5 stars, last pushed 3mo ago), licensed Apache-2.0. It adds 1,298 tokens to every session, about $0.0052 per session on Opus 5.5. 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.
oh-my-harness AGENTS.md compared with instructions from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| oh-my-harness AGENTS.mdguoxiaoshuai2023/oh-my-harness | Instructions file | 5 | 3mo ago | Apache-2.0 | A | 1,298 | Codex, OpenCode |
| vscode buildNext.instructions.mdmicrosoft/vscode | Instructions file | 195k | today | MIT | A | 9,719 | GitHub Copilot |
| next.js AGENTS.mdvercel/next.js | Instructions file | 143k | yesterday | MIT | A | 7,461 | Codex, OpenCode |
| vscode oss-third-party-notices.instructions.mdmicrosoft/vscode | Instructions file | 195k | today | MIT | A | 5,649 | GitHub Copilot |
| langchain AGENTS.mdlangchain-ai/langchain | Instructions file | 149k | today | MIT | A | 4,680 | Codex, OpenCode |
| vscode copilot-instructions.mdmicrosoft/vscode | Instructions file | 195k | today | MIT | A | 3,093 | GitHub Copilot |
| vscode ai-customization.instructions.mdmicrosoft/vscode | Instructions file | 195k | today | MIT | A | 1,907 | GitHub Copilot |