What is anatomy-of-a-knowledge-base AGENTS.md?

Repository instructions for working with a fictional company knowledge base that combines information from Confluence, JIRA, GitHub, and documents.

Running the local knowledge base, ingesting its source data, querying evidence through MCP or command-line tools, and changing the repository.

It gives coding agents the setup steps, operating rules, and fallback behavior they need to investigate the knowledge base correctly.

How gabrielrojasnyc/anatomy-of-a-knowledge-base configures Codex and O

Reuse

Borrowing it

Nothing to install: this file belongs to gabrielrojasnyc/anatomy-of-a-knowledge-base. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file

curl -O https://raw.githubusercontent.com/gabrielrojasnyc/anatomy-of-a-knowledge-base/main/AGENTS.md

Clone the repo

git clone --depth 1 https://github.com/gabrielrojasnyc/anatomy-of-a-knowledge-base

Made for: Codex, OpenCode.

Wrote 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.

README.md

[![agentmods](https://agentmods.dev/badge/instructions/gabrielrojasnyc/anatomy-of-a-knowledge-base/agents-md/github.svg)](https://agentmods.dev/instructions/gabrielrojasnyc/anatomy-of-a-knowledge-base/agents-md)

Your own site

<a href="https://agentmods.dev/instructions/gabrielrojasnyc/anatomy-of-a-knowledge-base/agents-md"><img src="https://agentmods.dev/badge/instructions/gabrielrojasnyc/anatomy-of-a-knowledge-base/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.

Your own site · 80×15

<a href="https://agentmods.dev/instructions/gabrielrojasnyc/anatomy-of-a-knowledge-base/agents-md"><img src="https://agentmods.dev/badge/instructions/gabrielrojasnyc/anatomy-of-a-knowledge-base/agents-md.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 1,171 This file is loaded in full into every session.

When invoked 1,171 The same file — it is already loaded in full.

Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.

Origin original No closer match found in the catalogue.

Token cost

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	$0.01171	$0.01171
Opus 5	$0.00585	$0.00585
Sonnet 5	$0.00234	$0.00234
Haiku 4.5	$0.00117	$0.00117

Measured 11d ago against content hash 9b41bd436acb, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-11, from the pricing page.

Security

Grade A, and why

anatomy-of-a-knowledge-base 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 11d 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.

AGENTS.md · 55 lines

How it starts

The opening of the file, as written. The whole thing — 55 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Agent operating notes

A runnable knowledge base over a fictional company, Helios. Four fixture sources (Confluence, JIRA, GitHub, a doc bucket) distilled into one Postgres table, queried by five parallel retrievers fused with RRF. You are the orchestrator; the MCP server only serves evidence. The full operating pattern, with a worked investigation, is docs/11-agent-playbook.md. To change the repo rather than query it, start at docs/12-changing-the-repo.md.

Setup, non-interactive

podman compose up -d (or docker compose up -d): Postgres with pgvector on port 5433.
pnpm install
pnpm kb init: green checkmarks when healthy. A missing CEREBRAS_API_KEY is a yellow warning, not a failure.
pnpm kb ingest: 5 to 45 minutes with a key (rate-limit pacing, not a hang), about 3 minutes without one.

Readiness: call the status tool (or pnpm kb search "checkpoint" --project helios-eng). An empty store means "not ingested", never "no evidence".

The MCP server

Claude Code discovers it from the committed .mcp.json; approve the prompt on first use. Other MCP clients launch it with pnpm --dir /path/to/repo kb-mcp over stdio. Eight tools, all LLM-free, with input schemas generated from the parameters the code actually reads.

Tool	Use when
`status`	anything looks empty or stale; separates "not ingested" from "no evidence"; no arguments
`search`	most questions; hybrid across all sources, takes `project` and `limit`
`get_document`	a result is worth reading whole; pass any result's `url` or a bare id like `HEL-482`
`search_confluence`, `search_jira`	you already know which system holds the answer
`search_code`	exact flags, error strings, function names; `re:` prefix for regex; a trailing `meta` row reports truncation
`who_knows`	routing a question to a person
`list_projects`	first call in a new session; takes no arguments

The investigation shape that works: search wide, get_document on the one or two urls worth reading in full, follow links into code, search_code for exact strings that are not links. Search returns ranked guesses; get returns the exact document a citation names. Url schemes: jira://HEL-482, confluence://HELIOS/HEL-008, bucket://file.md, github://helios/src/path.ts.

Read the full file on GitHub · 55 lines

Changes

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.

11d ago First seen · 55 lines · 1,171 tokens per session scan A 9b41bd436acb

Subscribe to this mod's changes

anatomy-of-a-knowledge-base AGENTS.md is an instructions file published in the GitHub repository gabrielrojasnyc/anatomy-of-a-knowledge-base (2 stars, last pushed 1mo ago), licensed MIT. It adds 1,171 tokens to every session, about $0.0059 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.

Building agents Databases

next.js AGENTS.md

AGENTS.md instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.

vercel/next.js · 7,296 tokens

codex AGENTS.md

AGENTS.md instructions for openai/codex, covering rust/codex-rs, the codex-core crate, code review rules, crate api surface and model visible context.

openai/codex · 5,153 tokens

vscode buildNext.instructions.md

Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).

microsoft/vscode · 6,785 tokens

anatomy-of-a-knowledge-base: Instructions file for Codex