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/litenova-solutions/fuse/agents-mdgit clone --depth 1 https://github.com/Litenova-Solutions/FuseWrote 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/litenova-solutions/fuse/agents-md)<a href="https://agentmods.dev/instructions/litenova-solutions/fuse/agents-md"><img src="https://agentmods.dev/badge/instructions/litenova-solutions/fuse/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 | $0.10172 | $0.10172 |
| Opus 5 | $0.05086 | $0.05086 |
| Sonnet 5 | $0.02034 | $0.02034 |
| Haiku 4.5 | $0.01017 | $0.01017 |
Grade A, and why
Fuse 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 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.
How it starts
The opening of the file, as written. The whole thing — 191 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Fuse - Agent and Contributor Guide
Read this before editing Fuse source or docs. It holds the durable context about the project so a session does not need it repeated in a prompt. For the full project briefing (architecture, all benchmark methodology and results, roadmap history), see briefing.md at the repo root.
What Fuse Is
Fuse is a .NET-native codebase context optimizer for AI-assisted development. It collects source files, reduces them for token efficiency, and emits one structured payload an agent or developer can read in a single call instead of opening thousands of files. It ships as a .NET global tool (fuse) and as a Model Context Protocol server (fuse mcp serve). It cuts tokens while keeping the public API intact, scopes to the files a task needs, and trims the round-trips an agent makes during its explore phase.
Repository Layout
src/Core: the pipeline and engine libraries -Fuse.Collection(discovery),Fuse.Reduction,Fuse.Emission,Fuse.Fusion(the orchestrator),Fuse.Indexing(the SQLite store and graph),Fuse.Semantics(workspace load, extraction, refactors),Fuse.Retrieval(ranking and resolution),Fuse.Scoping,Fuse.Context,Fuse.Workspace(resident compilation and test running).src/Host: the executable surfaces -Fuse.Cli(CLI commands, the MCP server, and thefuse hostdaemon) andFuse.BuildCaptureWorker(the out-of-process compiler-log worker).src/Plugins:Fuse.Plugins.Abstractions,Fuse.Plugins.Languages.CSharp,Fuse.Plugins.Languages.CSharp.Roslyn,Fuse.Plugins.Formats.Web.src/Packaging:Fuse.Capture, the build-target channel package (an MSBuild.targetsfile, no assembly).tests/: unit, golden-output, architecture, and integration tests.tests/benchmarks/: the harness, corpus manifest, and recorded results, in their own solution.site/: the documentation website (Next.js + Fumadocs), published at fuse.codes. All prose documentation lives here as MDX undersite/content/docs;site/publicholds the published static assets, including thefuse.jsonschema.assets/: the benchmark figure (fuse-benchmarks.pngand.svg) and the chart-generating script.mcp-registry/: the MCP Registry server manifest.briefing.md(repo root): full project briefing for planners (architecture, benchmarks, gaps); see alsoroadmap/for executable plans.- Solutions:
Fuse.slnxis the product;Fuse.Benchmarks.slnxholds the benchmark harness and is deliberately excluded from the product build and package.
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 · 191 lines · 10,172 tokens per session scan A dc51da0e1ab4
Fuse AGENTS.md is an instructions file published in the GitHub repository Litenova-Solutions/Fuse (34 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 10,172 tokens to every session, about $0.0509 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 2026-08-30.
Other instructions, from other repositories
roslyn-codelens-mcp CLAUDE.md
Claude Code instructions for MarcelRoozekrans/roslyn-codelens-mcp, covering roslyn codelens mcp, mcp server, skill, project structure and development.
graphify-dotnet copilot-instructions.md
Instructions for elbruno/graphify-dotnet, covering graphify-dotnet repository conventions, repository structure, ⚠️ critical rule: all code lives under src/, solution format and project conventions.
RoselineMCP CLAUDE.md
Instructions for Atypical-Consulting/RoselineMCP, covering claude.md, project overview, documentation alignment (required), architecture and service layer architecture.
aicodeprep-gui AGENTS.md
AGENTS.md instructions for detroittommy879/aicodeprep-gui, covering agents.md, maintaining this file, agent operational notes, package manager / running and licensing / pro access.
aicodeprep-gui CLAUDE.md
Claude Code instructions for detroittommy879/aicodeprep-gui: Read AGENTS.md which contains all the CLAUDE.md information.
DotNetAnalyzer CLAUDE.md
Claude Code instructions for CartapenaBark/DotNetAnalyzer, covering claude.md, 常用命令, 构建和测试, 打包和安装 and mcp 服务器.