mcpls rust-bridge.instructions.md

A set of engineering rules for a Rust bridge between an editor's code-intelligence protocol and an agent protocol. It covers typed positions, document caching, diagnostics, concurrency, and file watching.

In plain words
What is it for?
Use it when developing or reviewing the bridge's protocol types, document state, error handling, asynchronous locks, and file-watcher behavior.
Why use it?
It prevents subtle bugs such as mixing coordinate systems, reading stale files, missing protocol cases, or crashing on handler errors.

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

Made for: GitHub Copilot.

Per session 930 This file is loaded in full into every session.
When invoked 930 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 $0.00930 $0.00930
Opus 5 $0.00465 $0.00465
Sonnet 5 $0.00186 $0.00186
Haiku 4.5 $0.00093 $0.00093

Measured yesterday against content hash af6e2b799fc2, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

mcpls rust-bridge.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 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.

.github/instructions/rust-bridge.instructions.md · 81 lines

How it starts

The opening of the file, as written. The whole thing — 81 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Type safety

Position types (line, column) must be distinct newtypes, not bare u32. The MCP convention (1-based) and the LSP convention (0-based) must be expressed in the type system so that passing an MCP position directly to an LSP call is a compile error.

Protocol message variants must be an enum, not a string-matched dispatch. Matching on "textDocument/hover" strings is fragile; a closed enum of known methods makes unhandled cases visible at compile time and exhaustiveness-checked.

InboundMessage and similar enums that may gain variants in future protocol versions must be #[non_exhaustive] so downstream crates are not broken by additions.

Document state

Caching document content alongside a filesystem signature (mtime + size) prevents stale reads after external edits. On every access, stat the file and compare against the cached signature; re-read and re-notify the LSP server on mismatch.

Stat and read must use the same file handle to close the TOCTOU window. Open the file once with tokio::fs::File::open, call .metadata().await on the handle, then read through the same handle. A separate tokio::fs::metadata call before read_to_string leaves a window where an atomic rename can change the file between the two calls.

Document version numbers must be strictly monotone per URI. On overflow, reset to 1 rather than saturating — after didClose/didOpen the server treats the document as fresh and the version value is irrelevant to continuity.

Diagnostics pipeline

When a feature draws from multiple independent sources (pull request + push cache), fetch them concurrently with tokio::try_join! or tokio::join!. Awaiting them sequentially serialises work that has no data dependency.

A cache miss must be distinguishable from an empty result. Returning Vec::new() for both "no errors" and "not yet analysed" is ambiguous to callers. Use an Option<Vec> or a dedicated status type to express the difference.

Deduplication keys should prefer structured fields (error code) over free-form message text. When a code is present, key on (range, severity, code). The doc comment must describe the actual key — a mismatch between the comment and the implementation causes silent behavioural bugs.

Read the full file on GitHub · 81 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 · 81 lines · 930 tokens per session scan A af6e2b799fc2

Subscribe to this mod's changes

mcpls rust-bridge.instructions.md is an instructions file published in the GitHub repository bug-ops/mcpls (65 stars, last pushed 3d ago), licensed Apache-2.0. It adds 930 tokens to every session, about $0.0047 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.