capforge CLAUDE.md

capforge CLAUDE.md is an instructions file for coding agents from automatizatodo/capforge. It costs 629 tokens per session, scanned A, original, MIT.

Project instructions for CAPforge, an open-source server that helps AI agents work with SAP CAP/CDS and UI5 projects. SAP CAP is a framework for building cloud applications, while CDS is its data and service modeling language.

In plain words
What is it for?
They are for guiding changes to CAPforge, including its MCP tools, CAP and UI5 scaffolding, validation, linting, and deployment checks.
Why use it?
They give agents project-specific architecture, source, coding, security, and verification rules to follow.

Instructions file

Install

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.

agentmods
npx agentmods add instructions/automatizatodo/capforge/claude-md
Clone the repo
git clone --depth 1 https://github.com/automatizatodo/capforge

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.

agentmods badge for capforge CLAUDE.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/automatizatodo/capforge/claude-md.svg)](https://agentmods.dev/instructions/automatizatodo/capforge/claude-md)
Your own site
<a href="https://agentmods.dev/instructions/automatizatodo/capforge/claude-md"><img src="https://agentmods.dev/badge/instructions/automatizatodo/capforge/claude-md.svg" alt="Measured on agentmods" height="20"></a>
Per session 629 This file is loaded in full into every session.
When invoked 629 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
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.

ModelPer sessionOnce invoked
Fable 5 $0.00629 $0.00629
Opus 5 $0.00315 $0.00315
Sonnet 5 $0.00126 $0.00126
Haiku 4.5 $0.00063 $0.00063

Measured 4d ago against content hash 9c7e9a37a771, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

capforge CLAUDE.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 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.

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.

CLAUDE.md · 31 lines

How it starts

The opening of the file, as written. The whole thing — 31 lines — stays where its author put it; the contents beside it link to each section on GitHub.

CAPforge — project context

Open-source MCP server that makes AI agents (Claude Code, Cursor) get SAP CAP/CDS and UI5 right: modern scaffolding plus a real validation loop (cds compile).

Rules

  1. Never git push without explicit user confirmation.
  2. Commit messages: 4-5 words, imperative, no watermarks (no Co-Authored-By / Generated with).
  3. No comments in generated code; the code explains itself.
  4. Every tool change is verified before committing: node --check server.js plus the tests below.
  5. Files under knowledge/ are anchored to cited official sources — do not add rules without a source, or mark them as own additions (like the migration map).
  6. Examples use generic SAP canonical domains only (bookshop, ACME). Never real-world project data.

Architecture

  • server.js — the whole MCP server (stdio). Tools: cap_project_context, cap_scaffold_entity, cap_write_files, cap_validate, cap_lint, cap_deploy_check, cap_analyze_legacy_ui5, ui5_scaffold_section, ui5_validate_view, ui5_lint. Resource: conventions://cap (serves knowledge/ + conventions/ combined).
  • knowledge/ — canonical core (Free): CAP/UI5 best practices with sources. 01-modeling, 02-services, 03-annotations, 04-data-tooling, 05-migration (own map), 06-ui5.
  • conventions/ — house profile, overridable with env CAPFORGE_CONVENTIONS.
  • sandbox/ — minimal CAP project for the tests (includes ui5app/, the ui5lint testbed). docs/ — public website (GitHub Pages from /docs, domain capforge.dev).
  • extension/ — VS Code extension that registers the server via mcpServerDefinitionProviders (VS Code 1.101+). build.js bundles server.js with esbuild into server/server.mjs and copies knowledge/conventions; npx vsce package produces the .vsix. Setting capforge.conventionsPath -> env CAPFORGE_CONVENTIONS. No source duplication: the repo root is the single source of truth.
  • It calls no LLM (BYOK): the value is knowledge plus deterministic validation.

Verify

Read the full file on GitHub · 31 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.

  1. 4d ago First seen · 31 lines · 629 tokens per session scan A 9c7e9a37a771

Subscribe to this mod's changes

capforge CLAUDE.md is an instructions file published in the GitHub repository automatizatodo/capforge (0 stars, last pushed 25d ago), licensed MIT. It adds 629 tokens to every session, about $0.0031 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.