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/yanxreal/xcode-mpc/agents-mdgit clone --depth 1 https://github.com/YanxReal/Xcode-MPCWhat 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 | $0.02107 | $0.02107 |
| Opus 5 | $0.01053 | $0.01053 |
| Sonnet 5 | $0.00421 | $0.00421 |
| Haiku 4.5 | $0.00211 | $0.00211 |
Grade D, and why
Xcode-MPC AGENTS.md scanned grade D 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 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.
Reads agent configuration directoriesmediumAgent snooping
.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.
| **Codex** | `~/.codex/config.toml` → `[mcp_servers.xcode] command="node"` | `codex mcp list` | Recursive force deletehighDestructive command
rm -rf with a variable or a broad path is one typo away from removing the wrong tree.
- **Ask before destructive actions.** `make clean` / `rm -rf ~/Library/Developer/Xcode/DerivedData` / `git push --force` require explicit user confirmation. How it starts
The opening of the file, as written. The whole thing — 149 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS — How AI Should Work in This Repo
This is the single source of truth for any AI agent (OpenCode, Codex, Claude Code, Cursor, Copilot, etc.) working in this repository. If you are an AI, read this whole file before editing anything. If you are a human, this tells you what to expect from AI.
Version: 1.0.0 · Applies to: YanxReal/Xcode-MPC · Stack: index.js single-file MCP server, Yarn 4, Make, StdioServerTransport
1. Bilingual Rule — NON-NEGOTIABLE
This repo is bilingual by design:
| Language | Canonical files | Audience |
|---|---|---|
| English (primary) | README.md, docs/*.md |
International, GitHub default |
| Español (mirror) | README.es.md, docs/es/*.md |
Hispanohablante |
Rule: Every change to a user-facing Markdown file MUST be mirrored in both languages in the same commit.
- If you edit
README.md, you must editREADME.es.mdwith the equivalent Spanish translation. - If you edit
docs/installation.md, you must editdocs/es/installation.md. - If you edit
docs/tools.md, you must editdocs/es/tools.md, etc. - Do not leave one language behind. A PR that touches
docs/opencode.mdbut notdocs/es/opencode.mdwill be rejected.
How to translate:
- Keep structure identical (same headings, order, tables). Only language changes.
- Preserve code blocks, JSON/TOML,
shellEscapeexamples, andindex.js:linereferences verbatim — do not translate code. - Keep badges, links, and
{{placeholders}}intact. Translate only prose. - When in doubt, copy the English structure and translate paragraph by paragraph to Spanish.
Verification (run before commit):
# Both READMEs must have same tool count
grep -c "52 tools" README.md
grep -c "52 herramientas" README.es.md
# Both docs sets must exist
ls docs/*.md docs/es/*.md
# No 25/31/43/47 leftovers
grep -r "25 tools\|31 tools\|43 tools\|47 tools" --include="*.md" docs/ README* && echo "FAIL: stale count" || echo "OK"
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 · 149 lines · 2,107 tokens per session scan D 3b61d1a2b612
Xcode-MPC AGENTS.md is an instructions file published in the GitHub repository YanxReal/Xcode-MPC (1 stars, last pushed yesterday), licensed MIT. It adds 2,107 tokens to every session, about $0.0105 per session on Opus 5. A static security scan graded it D with 2 findings (reads agent configuration directories, recursive force delete). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other instructions, from other repositories
app-store-connect-mcp AGENTS.md
Instructions for erayendes/app-store-connect-mcp, covering agents.md, what this is, if a user asks you to install it, 1. agree on the profiles, then register them and 2. send the user to setup for the credentials.
the-unofficial-swift-programming-language-skill CLAUDE.md
Instructions for kylehughes/the-unofficial-swift-programming-language-skill, covering claude.md, repository purpose, build commands, generate the skill package and build options.
Apple-Agent-Kit CLAUDE.md
Claude Code instructions for caglarbaranbora/Apple-Agent-Kit, covering claude.md, layer order (do not violate), the normative specs, file naming and structure and validating.
Apple-Agent-Kit AGENTS.md
AGENTS.md instructions for caglarbaranbora/Apple-Agent-Kit, covering agents.md, startup procedure, rules, layer order and forbidden.
the-unofficial-swift-programming-language-skill AGENTS.md
Instructions for kylehughes/the-unofficial-swift-programming-language-skill, a project described as: The Swift Programming Language, packaged as a Skill for LLMs.
apple-dev-mcp CLAUDE.md
Instructions for tmaasen/apple-dev-mcp, covering claude.md, project overview, development commands, build and test and development.