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/jia-ethan/grok-keysmith/agents-mdgit clone --depth 1 https://github.com/Jia-Ethan/grok-keysmithWrote 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/jia-ethan/grok-keysmith/agents-md)<a href="https://agentmods.dev/instructions/jia-ethan/grok-keysmith/agents-md"><img src="https://agentmods.dev/badge/instructions/jia-ethan/grok-keysmith/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 5.1 | $0.00822 | $0.00822 |
| Opus 5 | $0.00411 | $0.00411 |
| Sonnet 5 | $0.00164 | $0.00164 |
| Haiku 4.5 | $0.00082 | $0.00082 |
Grade A, and why
grok-keysmith 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 6d 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 — 72 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Repository Agent Guide
Scope
These rules apply to the complete grok-keysmith repository. The project is a
zero-runtime-dependency Python CLI that manages global Grok Build instruction,
compatibility, hook, manifest, and recovery state under ~/.grok.
Release and External State
- Public repository history starts at
v0.1.1; do not import refs, commit IDs, tags, or metadata from the earlier private-only predecessor. - Keep existing public version tags unchanged: do not move, recreate, or overwrite them.
- Verify repository visibility, tags, Releases, assets, and settings live before documenting external state; a Git tag is not a GitHub Release.
- Creating a Release, workflow, repository setting, or other external publication action requires task-specific user approval.
Change Boundaries
- Preserve preview-first behavior. Deploy, uninstall, restore, and recovery
writes require explicit
--yes;--statusremains read-only. - Do not test write paths against the real
~/.grok. Use an isolated temporaryHOMEand retain failure evidence until the test completes. - Preserve fail-closed ownership checks for manifests, backups, hooks, config markers, journals, and immutable intent data.
- Do not patch Grok binaries, intercept network traffic, manage credentials, or alter running Grok processes.
- Treat
examples/grok-unrestricted.md,BUNDLED_PROMPT_B64, andBUNDLED_PROMPT_SHA256as one versioned unit. Any prompt change must update all three byte-for-byte and update README, CHANGELOG, and SECURITY hashes. - Keep Python 3.8 compatibility and standard-library-only runtime behavior unless a versioned compatibility change is explicitly approved.
- Keep the desktop icon aligned with the canonical Keysmith series source
(
gui/src-tauri/icons/source.png, SHA-256d7f5d09142b9bab6cd326fc18869e51d7adcba6d4f0929c703677186b7f3b347). Regenerate the complete Tauri icon set and syncgui/public/favicon.ico; do not replace only one platform asset.
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.
- 6d ago First seen · 72 lines · 822 tokens per session scan A ebc5471556b6
grok-keysmith AGENTS.md is an instructions file published in the GitHub repository Jia-Ethan/grok-keysmith (392 stars, last pushed 3d ago), licensed MIT. It adds 822 tokens to every session, about $0.0041 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 2026-08-30.
Other instructions, from other repositories
pi-xai-oauth AGENTS.md
AGENTS.md instructions for BlockedPath/pi-xai-oauth, covering agents.md — ai agent operations manual for pi-xai-oauth, project overview, key commands (exact, copy-paste ready), architecture & boundaries (must / must not) and file structure & wayfinding.
grokbuild-proxy CLAUDE.md
Instructions for GreyGunG/grokbuild-proxy, covering grokbuild, 项目状态, 架构, 开发与验证 and 变更约束.
goal-engineering AGENTS.md
Instructions for cobusgreyling/goal-engineering, covering agents.md — goal engineering reference, project purpose, verification, goal discipline and deny list.
grok-subagent AGENTS.md
Instructions for Walvez/grok-subagent, covering agents.md, 项目定位, 怎么跑起来, 依赖：node.js 22+、已登录的官方 grok cli（/.grok/bin/grok 或 grok） and 安装后必须新建 codex 任务，旧会话不会热加载 skill/mcp.
grok-codex-plugin AGENTS.md
AGENTS.md instructions for jinxlzc/grok-codex-plugin: Grok Build plugin that commands the local Codex CLI.
grok-build-showcase AGENTS.md
Instructions for cobusgreyling/grok-build-showcase, covering agents.md — grok build rules for grok-build-showcase, coding standards, build, test & verification, git & version control and architecture notes.