mcpls mcp-tools.instructions.md

A set of rules for adding MCP tools to a Rust server. MCP is a protocol that lets an AI agent discover and call external tools.

In plain words
What is it for?
Use it when defining tool parameters, registering tools, documenting absolute paths and one-based positions, handling errors, and implementing protocol dispatch.
Why use it?
It reduces failures caused by missing schemas, incorrect path or position formats, panics, and vague error messages.

Instructions file for GitHub Copilot

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/bug-ops/mcpls/mcp-tools
Clone the repo
git clone --depth 1 https://github.com/bug-ops/mcpls

Made for: GitHub Copilot.

Per session 357 This file is loaded in full into every session.
When invoked 357 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 $0.00357 $0.00357
Opus 5 $0.00179 $0.00179
Sonnet 5 $0.00071 $0.00071
Haiku 4.5 $0.00036 $0.00036

Measured 2d ago against content hash 03ee734460ca, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

mcpls mcp-tools.instructions.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 2d 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.

.github/instructions/mcp-tools.instructions.md · 39 lines

What it actually says

Type safety

Tool parameter structs must use domain types, not raw primitives, for values with protocol-specific conventions. Position fields (line, column) must be documented as 1-based in both the /// doc comment and the #[schemars(description)] — the MCP↔LSP coordinate conversion happens in the bridge layer, not here.

Path parameters must accept absolute paths. State this in the schema description; AI clients use it to form requests correctly.

New tool structs require #[derive(JsonSchema)]. Omitting it silently removes the tool from the schema exposed to clients — they cannot discover or invoke it.

Error handling

Tool handlers must return structured MCP error responses on failure, never panic. unwrap() and expect() in handler code kill the entire server process and drop all active client sessions. Use ? with proper error mapping instead.

All error variants must carry enough context for the caller to understand what went wrong. A bare "internal error" string is not actionable. Include the file path, method name, or other relevant detail.

Idiomatic dispatch

When adding a new tool, register it in both the capability advertisement (tool list) and the dispatch match arm. A tool missing from the list is undiscoverable; a tool missing from dispatch panics at runtime. A compile-time check (e.g. a const assertion or a test that calls every listed tool name through dispatch) is preferable to relying on manual review.

Prefer exhaustive match arms over wildcard catch-alls in dispatch. A wildcard silently swallows unhandled tool names; exhaustive matching makes unregistered names a compile error.

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. 2d ago First seen · 39 lines · 357 tokens per session scan A 03ee734460ca

Subscribe to this mod's changes

mcpls mcp-tools.instructions.md is an instructions file published in the GitHub repository bug-ops/mcpls (66 stars, last pushed 4d ago), licensed Apache-2.0. It adds 357 tokens to every session, about $0.0018 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.