jeremylongshore/claude-code-slack-channel — Two-way Slack bridge for Claude Code.

Description and use

**Brief summary**
The *claude‑code‑slack‑channel* project provides a two‑way Slack bridge for Claude Code, using Slack’s Socket Mode and the Model Context Protocol (MCP) over stdio. It isolates sessions per thread, records every interaction in a hash‑chained tamper‑evident audit journal, and enforces a five‑layer prompt‑injection defense, while exposing policy‑gated MCP tools through Block‑Kit button relays. The code runs on Bun, Node.js, or Docker, is written in strict TypeScript, and ships with a ready‑to‑use CLI/SDK.

**Value**

• **Secure, auditable AI‑tool integration** – the tamper‑evident journal and multi‑layer injection defenses let organizations let Claude Code act on real data and internal tools without fearing silent misuse.
• **Standardized protocol** – by speaking MCP, the bridge can be swapped for any other MCP‑compatible service, making it a reusable “AI‑to‑Slack” adapter across projects.
• **Operational flexibility** – three runtimes (Bun, Node, Docker) let teams run the bridge in development, CI pipelines, or production containers with minimal friction.

**Practical adoption path**

1. **Prototype** – clone the repo, run the Docker compose or bun run dev to connect a test Slack workspace and a Claude Code instance.
2. **Configure policies** – define MCP tool permissions and prompt‑injection rules in the provided JSON/YAML files; use the Block‑Kit buttons to test permission relays.
3. **Integrate** – replace the test Claude endpoint with the production model, point the bridge at the target Slack channel, and enable the audit‑journal sink (e.g., write to S3 or a SIEM).
4. **Deploy** – package the bridge as a Docker image or a serverless function, embed it in your CI/CD pipeline, and monitor health via the built‑in metrics endpoint.

**Production readiness**

• **Activity & community** – 21 ★, 16 forks, recent commits (last updated 2026‑05‑14) and multiple runtimes indicate an active code base.
• **Security posture** – tamper‑evident journaling and explicit policy gating reduce attack surface; however, a formal security audit and license verification are still required.
• **Reliability** – strict TypeScript typing, socket‑mode reconnection logic, and Docker‑compatible builds give confidence for pilot deployments; scaling to many channels will need load‑testing.

Overall, the project shows strong signals for a serious pilot: clear security controls, straightforward integration steps, and solid engineering practices, pending final license and security reviews.

Problem

jeremylongshore/claude-code-slack-channel helps connect AI assistants to real tools and data through a standard protocol.

Outcome

High for an OSS candidate: recent activity, adoption, and ecosystem signals are strong enough for a serious pilot.

Workflow

• connect AI agents to tools
• ship Model Context Protocol servers
• standardize integrations

Evidence

Two-way Slack bridge for Claude Code. Socket Mode + MCP stdio. Hash-chained tamper-evident audit journal, per-thread session isolation, policy-gated MCP tools, permission relay with Block Kit buttons, five-layer prompt-injection defense. Three runtimes: Bun, Node.js, Docker. TypeScript strict.

Source