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/dimosgit/grounded-knowledge-engine/agents-mdgit clone --depth 1 https://github.com/dimosgit/grounded-knowledge-engineWhat 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 | $0.02449 | $0.02449 |
| Opus 5 | $0.01224 | $0.01224 |
| Sonnet 5 | $0.00490 | $0.00490 |
| Haiku 4.5 | $0.00245 | $0.00245 |
Grade A, and why
grounded-knowledge-engine 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 yesterday.
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 — 190 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Agent Guide
Canonical guidance for coding agents (Claude Code, Codex, Gemini CLI) working
in this repository. CLAUDE.md and GEMINI.md import this file — edit here,
not there.
GKE is a local-first grounded knowledge engine: Markdown files are the source of truth, the retrieval index (BM25 / SQLite FTS5) is derived and disposable, and the same engine core is exposed through a CLI, a local MCP server, and the optional Operator Cockpit web preview. This repo is public — see Sanitization below before committing anything.
Agent operating contract
This file is the repository's agent contract. Do not create a second contract file with overlapping rules.
- Use the connected GKE MCP server first when a request depends on the user's
documents, a named workspace or client, previous research, project state,
decisions, or handoffs. For ordinary grounded Q&A use
kb.answer_and_capture; for a named project usekb.resume_project; usekb.searchonly for evidence-only retrieval andkb.get_recordonly for an explicitly requested record. - Do not use GKE MCP merely to inspect or debug this repository's source code. The checked-out files and tests are authoritative for implementation work.
- If a task should use GKE but no
kb.*tools are visible, say that the MCP connection is unavailable, runnpm run setup:mcp -- --client <client>when local configuration is in scope, and remind the operator that clients must be restarted to reload their tool catalog. Do not silently pretend an ungrounded answer came from GKE. - MCP tools are model-controlled: registration makes them available but does not force a call. These routing rules are therefore mandatory for applicable tasks.
Authorship and attribution
AI assistants are tools, not repository contributors. Never use an AI product,
model, bot, or session identity as a commit author, committer, co-author,
reviewer signature, or release author. Do not add AI attribution trailers such
as Co-Authored-By or session links. Preserve the human operator's configured
Git identity for commits unless the operator explicitly requests another human
identity.
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.
- yesterday First seen · 190 lines · 2,449 tokens per session scan A ffec3f28d1e4
grounded-knowledge-engine AGENTS.md is an instructions file published in the GitHub repository dimosgit/grounded-knowledge-engine (5 stars, last pushed 5d ago), licensed MIT. It adds 2,449 tokens to every session, about $0.0122 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
siyuan AGENTS.md
AGENTS.md instructions for siyuan-note/siyuan, covering agents.md, 1. non-negotiable constraints, do not hand-edit, verification and prohibited operations and 2. project-specific rules.
obsidian-mcp-server AGENTS.md
Instructions for cyanheads/obsidian-mcp-server, covering agent protocol, what's next?, core rules, patterns and tool — obsidianlisttags.
Starcat AGENTS.md
Instructions for starcat-app/Starcat, covering agents.md, 🚨 硬性铁律（每次写代码前必读，违反即返工）, 🌿 git 分支与 worktree（强制）, 🧭 主进度索引（每次开工前必读） and 状态符号（与功能实现总览.md 同步）.
Starcat CLAUDE.md
Instructions for starcat-app/Starcat, covering claude.md, 🚨 硬性铁律（每次写代码前必读，违反即返工）, 🌿 git 分支与 worktree（强制）, 🧭 主进度索引（每次开工前必读） and 状态符号（与功能实现总览.md 同步）.
Vault-Agent-Memory AGENTS.md
Instructions for zycaskevin/Vault-Agent-Memory, covering agent instructions for vault agent memory, common install architecture, second decision: optional features, first decision: database scope and safe agent workflow.
my-wiki AGENTS.md
Instructions for NimaChu/my-wiki, covering my wiki agent project, project map, knowledge model, route requests and development checks.