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/thangchung/agent-engineering-experiment/copilot-instructionsgit clone --depth 1 https://github.com/thangchung/agent-engineering-experimentWhat 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 | $0.01551 | $0.01551 |
| Opus 5 | $0.00776 | $0.00776 |
| Sonnet 5 | $0.00310 | $0.00310 |
| Haiku 4.5 | $0.00155 | $0.00155 |
Grade A, and why
agent-engineering-experiment copilot-instructions.md scanned grade A 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 3d 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.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
| `ExecTool.cs` | Shell exec with safety blocklist (rm, sudo, curl, etc.) | How it starts
The opening of the file, as written. The whole thing — 119 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Copilot Instructions
Principles
- KISS / YAGNI / DRY / Boy-Scout. Simplest thing that works.
- Red-green TDD whenever possible.
Build & Run
dotnet build dotnetclaw.slnx # build all projects
dotnet run --project DotNetClaw/DotNetClaw.csproj # standalone
dotnet run --project DotNetClaw.AppHost/DotNetClaw.AppHost.csproj # Aspire (coffeeshop-cli + DotNetClaw)
Secrets: dotnet user-secrets set "Slack:BotToken" "xoxb-..." etc. See appsettings.json for all keys.
Solution Structure
DotNetClaw/ Main ASP.NET Core app (net10.0)
DotNetClaw.AppHost/ Aspire orchestration (local dev, registers coffeeshop-cli + dotnetclaw)
DotNetClaw.ServiceDefaults/ Shared Aspire defaults (health, OTel, resilience)
DotNetClaw.Tests/ Integration tests (console runner)
Architecture
Slack user Browser user
↓ Socket Mode ↓ HTTP POST /api/chat
SlackChannel.cs WebChannel.cs
└──────────┬───────┘
↓ sessionId + text
ClawRuntime.cs
ConcurrentDictionary<sessionId, AgentSession>
↓
AIAgent (MAF 1.0 + GitHub Copilot SDK 0.2.1)
ReAct loop + streaming + tool calling
↓
CopilotClient (spawns copilot-cli via stdio JSON-RPC)
↓
mind/ directory → system message
Session IDs — one AgentSession per unique ID:
- Slack DM:
slack:dm:{userId}, Slack mention:slack:{channelId}:{userId} - Web:
web:{cookieGuid}
File Map
| File | Role |
|---|---|
Program.cs |
Composition root — DI, AIAgent factory, channel wiring, endpoints |
ClawRuntime.cs |
Session store, per-session semaphore, HandleAsync / HandleStreamingAsync |
SlackChannel.cs |
Slack Socket Mode — DM + @mention handlers, extension methods |
WebChannel.cs |
HTTP chat API + static file serving, cookie-based sessions |
MindLoader.cs |
Loads mind/SOUL.md + agent files + working-memory → system message |
MemoryTool.cs |
Agent tools: AppendLog, AddRule, SaveFact (writes to mind/.working-memory/) |
ExecTool.cs |
Shell exec with safety blocklist (rm, sudo, curl, etc.) |
SkillLoaderTool.cs |
Skills via CLI or MCP from coffeeshop-cli (ListSkills, LoadSkill, ListMenuItems, LookupCustomer, SubmitOrder) |
wwwroot/index.html |
Browser chat UI (vanilla HTML + JS, no framework) |
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.
- 3d ago First seen · 119 lines · 1,551 tokens per session scan A 7a3c0b0124d2
agent-engineering-experiment copilot-instructions.md is an instructions file published in the GitHub repository thangchung/agent-engineering-experiment (24 stars, last pushed 1mo ago), licensed MIT. It adds 1,551 tokens to every session, about $0.0078 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
foundry-local cs-debug-binding.instructions.md
Use when changing the C ABI (foundrylocalc.h) and validating C# tests, or when C# tests fail with stale-ABI symptoms (struct/layout mismatches, dispatch falling through unexpectedly).
maui integration-tests.instructions.md
Guidance for GitHub Copilot when working with .NET MAUI integration tests.
maui handler-patterns.instructions.md
Instructions for dotnet/maui, covering handler mapper and property patterns, property update flow, lifecycle management, null safety in callbacks and native defaults preservation.
maui performance-hotpaths.instructions.md
Instructions for dotnet/maui, covering performance-critical path rules, hot paths in maui, allocation avoidance, caching and invalidation and collection iteration.
cryptoexchange-net
Use CryptoExchange.Net abstractions when generating C#/.NET code that needs to work across MULTIPLE cryptocurrency exchanges (Binance + Bybit + OKX + Kraken + Coinbase + ...) — for arbitrage, best-execution routing, multi-exchange portfolio dashboards, exchange-agnostic trading bots, or comparison tools. Triggers on…
Kucoin.Net copilot-instructions.md
Instructions for JKorf/Kucoin.Net, covering copilot instructions for kucoin.net, use kucoin.net, not raw http, client setup, result handling and api structure.