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/mrf/godot-stagehand/agents-mdgit clone --depth 1 https://github.com/mrf/godot-stagehandWrote 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/mrf/godot-stagehand/agents-md)<a href="https://agentmods.dev/instructions/mrf/godot-stagehand/agents-md"><img src="https://agentmods.dev/badge/instructions/mrf/godot-stagehand/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 | $0.01985 | $0.01985 |
| Opus 5 | $0.00992 | $0.00992 |
| Sonnet 5 | $0.00397 | $0.00397 |
| Haiku 4.5 | $0.00198 | $0.00198 |
Grade C, and why
godot-stagehand AGENTS.md scanned grade C 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 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.
Recursive force deletehighDestructive command
rm -rf with a variable or a broad path is one typo away from removing the wrong tree.
rm -rf directory # NOT: rm -r directory How it starts
The opening of the file, as written. The whole thing — 137 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Agent Instructions
This project uses br (beads_rust) for issue tracking. Run br robot-docs guide to get started.
Quality Standards — MANDATORY
Every agent session MUST follow strict TDD. No exceptions.
TDD Protocol
- Write a failing test FIRST — before any implementation code
- Go: unit test in
*_test.goor E2E test ine2e_test.go - GDScript: GdUnit4 test in
testdata/test_project/test/
- Go: unit test in
- Run the test — verify it fails for the right reason
- Write the minimum code to make the test pass
- Run the full test suite —
go test ./...— no regressions - Refactor if needed, re-run tests
- Run
code-simplifier:code-simplifieragent before committing
Quality Gates (run before every commit)
go vet ./... # Lint
go test ./... # All Go tests pass
# If GDScript changed — run the GdUnit4 unit suite (headless):
GODOT_BIN=/path/to/Godot_v4.x-stable_linux.x86_64 ./scripts/run-gdscript-tests.sh
# Or through Go, which parses the JUnit report and fails on any failure/skip:
GODOT_BIN=/path/to/Godot_v4.x-stable_linux.x86_64 go test -tags=gdscript -run TestGdUnitSuite .
# Needs Godot 4.6+: GdUnit4 v6.1.3's CLI runner hangs on 4.4 headless. The addon
# itself still targets 4.3+, which the gdscript-parse CI matrix covers.
# If GDScript changed — strict-mode validation:
timeout 5 ${GODOT_BIN:-godot} --path testdata/test_project --headless --stagehand 2>&1 | grep -E "SCRIPT ERROR|Server listening"
# Must show "Server listening on port 26700" with ZERO "SCRIPT ERROR" lines.
# testdata/test_project has all GDScript warnings elevated to errors — this catches:
# - Implicitly inferred static types (:= without annotation)
# - float()/int()/bool()/String() constructors with Variant args
# - Discarded return values
# - Variable shadowing
# - Untyped function parameters
# - Unsafe property/method access on Variant
# Set GODOT_BIN to your Godot binary (e.g. export GODOT_BIN=/path/to/Godot_v4.x-stable_linux.x86_64)
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 · 137 lines · 1,985 tokens per session scan C c1956e36d04f
godot-stagehand AGENTS.md is an instructions file published in the GitHub repository mrf/godot-stagehand (7 stars, last pushed 4d ago), licensed MIT. It adds 1,985 tokens to every session, about $0.0099 per session on Opus 5. A static security scan graded it C with 1 finding (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
godot-docs-mcp CLAUDE.md
Claude Code instructions for mcintalmo/godot-docs-mcp, covering godot-docs-mcp — working notes, non-negotiables, layer map, conventions and adding a tool.
Godot-MCP CLAUDE.md
Instructions for IvanMurzak/Godot-MCP, covering claude.md — godot-mcp, tool families, build / test, suite 1 — build (ci gate; 0 errors required) and cli (godot-cli).
yume CLAUDE.md
Instructions for kamwoh/yume, covering yume (夢) — json-driven game framework on godot, quick start, framework structure, how the engine works and tick rate is the engine's heartbeat (2026-05-16).
wazzicode-godot AGENTS.md
AGENTS.md instructions for robertobendi/wazzicode-godot, a project described as: WazziCode for Godot — production-grade local MCP tools, editor automation, diagnostics, playtesting, and agent workflows.
wazzicode-godot CLAUDE.md
Claude Code instructions for robertobendi/wazzicode-godot: This is a pnpm TypeScript workspace plus a Godot 4 editor addon and a Tauri desktop app.
comedot AGENTS.md
AGENTS.md instructions for InvadingOctopus/comedot, covering agents.md, guidelines for the comedot project and repository, project overview, folder structure & module organization and subsystems.