SharpClawCode CLAUDE.md

Project instructions for SharpClaw Code, a C# and .NET coding-agent runtime. They document the repository, its main program flow, and commands for building, testing, and running it.

In plain words
What is it for?
Use them to understand the repository architecture, build the solution, run all or selected tests, run parity tests, and launch the command-line application.
Why use it?
They give coding agents the project's conventions and reliable commands, reducing guesswork when changing or checking the code.

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/clawdotnet/sharpclawcode/claude-md
Clone the repo
git clone --depth 1 https://github.com/clawdotnet/SharpClawCode
Per session 1,017 This file is loaded in full into every session.
When invoked 1,017 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 $0.01017 $0.01017
Opus 5 $0.00508 $0.00508
Sonnet 5 $0.00203 $0.00203
Haiku 4.5 $0.00102 $0.00102

Measured yesterday against content hash ea0abcbe9197, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

SharpClawCode 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 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.

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 · 86 lines

How it starts

The opening of the file, as written. The whole thing — 86 lines — stays where its author put it; the contents beside it link to each section on GitHub.

CLAUDE.md

This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.

Project

SharpClaw Code is a C#-native coding-agent runtime (.NET 10 / C# 13) inspired by the Rust claw-code surface. It provides durable sessions, permission-aware tool execution, provider abstraction (Anthropic & OpenAI-compatible), MCP integration, and plugin/skill extensibility.

Build & Test Commands

# Build
dotnet build SharpClawCode.sln

# Run all tests
dotnet test SharpClawCode.sln

# Run a single test by name
dotnet test SharpClawCode.sln --filter "FullyQualifiedName~YourTestName"

# Run parity harness scenarios only
dotnet test SharpClawCode.sln --filter "FullyQualifiedName~ParityScenarioTests"

# Run the CLI
dotnet run --project src/SharpClaw.Code.Cli -- [arguments]

Architecture

Entry point: src/SharpClaw.Code.Cli/Program.cs builds a host via CliHostBuilder, resolves CliCommandFactory, and invokes System.CommandLine.

Core execution flow:

  1. ConversationRuntime.RunPromptAsync resolves/creates a session
  2. DefaultTurnRunner builds prompt context via IPromptContextAssembler
  3. PrimaryCodingAgent delegates to AgentFrameworkBridge -> ProviderBackedAgentKernel -> IModelProvider
  4. Tools execute through IToolRegistry -> PermissionPolicyEngine -> ISharpClawTool.ExecuteAsync
  5. Events published via IRuntimeEventPublisher, sessions persisted as snapshots + NDJSON append logs

DI composition roots: RuntimeServiceCollectionExtensions.AddSharpClawRuntime() and CliServiceCollectionExtensions.AddSharpClawCli().

Key project boundaries:

  • Protocol is the dependency-light center: DTOs, enums, events, ProtocolJsonContext (source-gen JSON)
  • Runtime owns orchestration and lifecycle, not the CLI
  • Cli and Commands are thin; business logic lives in Runtime/Agents/Tools
  • Tools must route through Permissions -- never bypass permission checks
  • Sessions supports both durable snapshots and append-only event logs
  • Agents wraps the Microsoft Agent Framework without leaking provider/transport details

Read the full file on GitHub · 86 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. yesterday First seen · 86 lines · 1,017 tokens per session scan A ea0abcbe9197

Subscribe to this mod's changes

SharpClawCode CLAUDE.md is an instructions file published in the GitHub repository clawdotnet/SharpClawCode (85 stars, last pushed 1mo ago), licensed MIT. It adds 1,017 tokens to every session, about $0.0051 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.