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/abap2ui5/mcp-server/agents-mdgit clone --depth 1 https://github.com/abap2UI5/mcp-serverWhat 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 | $0.05770 | $0.05770 |
| Opus 5 | $0.02885 | $0.02885 |
| Sonnet 5 | $0.01154 | $0.01154 |
| Haiku 4.5 | $0.00577 | $0.00577 |
Grade A, and why
mcp-server AGENTS.md scanned grade A with 1 finding 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 2d 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.
Runs shell commandslowCapability
Expected in a hook, worth knowing in a rule or an instructions file.
const { spawn } = require("child_process"); How it starts
The opening of the file, as written. The whole thing — 320 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md — mcp-server
Single source of truth for agents working on the abap2UI5 MCP server —
the app_guide → validate_view/screenshot_view → deploy_app → build_backend → run_app loop exposed to MCP clients, no SAP system required.
The loop has a cheap half and an expensive half, and keeping them apart is
the point. validate_view and screenshot_view both work from SOURCE
through the linter's render harness: seconds, no backend, no transpile, and
blind to everything that only exists at runtime. build_backend/run_app
boot the real transpiled app: tens of minutes for the first build, and the
only place the ABAP actually runs. A tool that moves work from the second half
to the first is worth more here than almost anything else — the expensive half
is what an agent's feedback loop is made of.
This entire project is in English. No non-ASCII literal goes in a source file — the CAPABILITIES.md status marks are built with
String.fromCodePointfor exactly this reason (lib/capabilities.mjs), so a parser's data can never depend on how an editor saved a glyph. Prose is a different matter: comments and the tool descriptions have always used em dashes and stay as they are. (This paragraph said "source files are 7-bit ASCII", which no file in the repo has ever been,lib/capabilities.mjsincluded.)test/ascii.test.mjsis the gate, overserver.mjsand everylib/*.mjs: everything above U+007F fails except the prose punctuation listed there by code point (em dash, en dash, section sign), each with the reason it is allowed. An emoji, an invisible character or a homoglyph letter is refused by file and line — build it from its code point instead, the waylib/capabilities.mjsbuilds the status marks.
The one thing to understand first: this repo cannot work alone
mcp-server bundles no content. Every tool reads live from sibling checkouts,
resolved per call in lib/repos.mjs (explicit env var, then the ../<name>
sibling of this repo — plus, for abap2UI5, the in-repo .abap2UI5 clone that
samples-controls' npm run node:setup creates). A set env var is
authoritative: when it points at a directory without the expected checkout,
the repo resolves to null and the tool reports the misconfiguration — there
is no silent fallback to the sibling guess.
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.
- 2d ago First seen · 320 lines · 5,770 tokens per session scan A 60ea9bce6912
mcp-server AGENTS.md is an instructions file published in the GitHub repository abap2UI5/mcp-server (1 stars, last pushed 2d ago), licensed MIT. It adds 5,770 tokens to every session, about $0.0289 per session on Opus 5. A static security scan graded it A with 1 finding (runs shell commands). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other instructions, from other repositories
abap2UI5 AGENTS.md
Instructions for abap2UI5/abap2UI5, covering agents.md — ai assistant guide for abap2ui5, where knowledge lives, project overview, related repositories and architecture.
abap2UI5 CLAUDE.md
Instructions for abap2UI5/abap2UI5: All project guidance lives in AGENTS.md — the single source of truth for this repository (architecture and the layered design, the src/00-99 packages, the generated app/webapp mirror, the build and delivery pipeline, the CI gates and ABAP code style).
samples CLAUDE.md
Instructions for abap2UI5/samples: All project guidance now lives in AGENTS.md — the single source of truth for this repository (folder scheme, src/01 vs src/00 compatibility model, overview generation rules, ABAP code style, and app structure).
samples AGENTS.md
Instructions for abap2UI5/samples, covering agents.md, language, pull requests, 1. repository layout and the restricted and obsolete packages are gone.
DevoxxGenieIDEAPlugin GEMINI.md
Instructions for devoxx/DevoxxGenieIDEAPlugin, a project described as: DevoxxGenie is an agentic plugin for IntelliJ IDEA that uses local LLM's (Ollama, LMStudio, GPT4All, Jan and Llama.cpp) and Cloud based LLMs to help review, test, explain your project code. Latest version now also supports Spec Driven Development…
vscode_abap_remote_fs CLAUDE.md
Instructions for marcellourbani/vscode_abap_remote_fs, covering abap fs development guidelines, workflow and constraints.