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/zach-blumenfeld/knowledge-index/agents-mdgit clone --depth 1 https://github.com/zach-blumenfeld/knowledge-indexWrote 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/zach-blumenfeld/knowledge-index/agents-md)<a href="https://agentmods.dev/instructions/zach-blumenfeld/knowledge-index/agents-md"><img src="https://agentmods.dev/badge/instructions/zach-blumenfeld/knowledge-index/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 | $0.03921 | $0.03921 |
| Opus 5 | $0.01961 | $0.01961 |
| Sonnet 5 | $0.00784 | $0.00784 |
| Haiku 4.5 | $0.00392 | $0.00392 |
Grade C, and why
knowledge-index AGENTS.md scanned grade C with 2 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 4d 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.
Downloads and executes remote codehighSupply chain
curl | sh runs whatever the server returns today, which is not necessarily what it returned when this was reviewed.
| `install.sh` | One-command installer (`curl -sSfL https://knowledge-index.ai/install.sh \| bash`): uv → `ki` → `neo4j-cli` → agent skills. Idempotent. Served verbatim at the install URL. | Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
| `install.sh` | One-command installer (`curl -sSfL https://knowledge-index.ai/install.sh \| bash`): uv → `ki` → `neo4j-cli` → agent skills. Idempotent. Served verbatim at the install URL. | How it starts
The opening of the file, as written. The whole thing — 136 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
Instructions for AI agents (Claude, Codex, Cursor, etc.) working on the knowledge-index codebase. If you are an agent being asked by a user to use ki (index their notes, search their knowledge base, etc.), read skills/knowledge-base/SKILL.md instead.
What this repo is
knowledge-index (CLI: ki) is a search index over a folder of markdown, backed by Neo4j — it syncs the filesystem into a knowledge graph and serves fast search / navigation / retrieval. The primitive verbs are ki index (sync) and ki search (query); read/navigation (ki outline [alias ki tree], ki get, ki status, ki vault list), index removal (ki drop, ki nuke), and setup (ki configure, ki profile list, ki init, ki skill) sit on top.
Canonical design lives in docs/scoping.md (profiles, vaults, config, the command surface, and the local/remote scoping model) and docs/general-philosophy.md (the tenets); the schema is docs/data-model/schema.md. docs/README.md indexes everything.
Non-negotiable design principles
These constrain every change you make. If a proposed feature violates one of these, reject the proposal rather than working around it.
kiis an index, not a document store. Never mutate user-owned source files (.md). Allkioutput lives in~/.config/ki/(config), Neo4j (the index), or.ki/vault.yaml(vault identity + optional user-authored description per vault).kiwrites theuri:field on first creation and otherwise touches user-authored fields only when the user explicitly asks via a flag (ki index --description "..."writesdescription:; without the flag,kiis read-only). No--purgeflag, no "auto-organize my notes," no rewriting frontmatter. Seedocs/general-philosophy.md.- The backend is opaque to the user. From the user's and the agent-as-user's perspective,
kiis a search tool. They don't need to know about Cypher, Neo4j, or graph traversal. Don't surface backend concepts (Cypher errors, node labels, etc.) in default output. - One source of truth per concern. Config lives at
~/.config/ki/config.yaml; vault identity + per-vault user metadata live in.ki/vault.yaml; graph data lives in Neo4j. Don't introduce parallel state. - Safe by default, dangerous by flag. Destructive operations (whole-vault removal) require explicit flags AND typed confirmation. Cloud-resource creation (Aura) requires explicit consent even on agent auto-mode. See
docs/general-philosophy.md(safe by default) andskills/knowledge-base/SKILL.md(agent auto-mode rules).
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.
- 4d ago First seen · 136 lines · 3,921 tokens per session scan C 7c6e7a426681
knowledge-index AGENTS.md is an instructions file published in the GitHub repository zach-blumenfeld/knowledge-index (6 stars, last pushed 10d ago), licensed Apache-2.0. It adds 3,921 tokens to every session, about $0.0196 per session on Opus 5. A static security scan graded it C with 2 findings (downloads and executes remote code, makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other instructions, from other repositories
memory AGENTS.md
Instructions for sunbeamdotpt/memory, covering sunbeam-memory — agent guide, project overview, runtime modes, semantic memory search (optional) and technology stack.
aeon CLAUDE.md
Instructions for aeonfun/aeon, covering aeon, how aeon works, strategy, voice and soul file hierarchy (read in this order).
cognirepo CLAUDE.md
Claude Code instructions for ashlesh-t/cognirepo, covering claude.md, key rules, session start sequence (run in this order), behavioral confirmation rule and personas (cognirepo-402, cognirepo-403).
wayland-core copilot-instructions.md
Copilot instructions for FerroxLabs/wayland-core, covering ijfw rules, output discipline, memory routing, context discipline and cross-audit.
mcp-structured-memory CLAUDE.md
Claude Code instructions for nmeierpolys/mcp-structured-memory, a project described as: Structured Memory MCP Server.
inkwell-memory CLAUDE.md
Instructions for veronchenko/inkwell-memory, covering claude.md — inkwellmemory, layout, multi-tenant mode (inkwellmultitenant=1), conventions and testing.