- Home
- Skills
- Agents & Orchestration
- Sluice: Allowlisted MCP Tools for Coding Agents
Works with the AI Agents you already use
Sluice: Allowlisted MCP Tools for Coding Agents
A secure wrapper for agents to interact with GitHub, Vercel, and Supabase via a deny-by-default whitelist.
$5
Sluice: Allowlisted MCP Tools for Coding Agents
Example session with this skill installed
Check if there are any open critical issues in the github server and summarize the recent sentry errors for the same project.
- Read your context and instructions
- Compiled the sluice
I have queried the GitHub and Sentry servers via sluice. I retrieved 3 critical issues from GitHub and 5 recent error events from Sentry.
Project Status Summary
| Service | Category | Description | Status |
|---|---|---|---|
| GitHub | Issue #402 | Critical: Buffer overflow in parser | Open |
| GitHub | Issue #398 | Urgent: Auth provider timeout | Open |
| GitHub | Issue #385 | Security: Dependency vulnerability CVE-2024-X | Open |
| Sentry | Error | TypeError: null is not an object in /api/auth | 12 events |
| Sentry | Error | TimeoutError in Database connection pool | 45 events |
| Sentry | Error | RateLimitExceeded on GitHub API integration | 8 events |
Next steps
- Would you like me to link the Sentry
TimeoutErrorto GitHub Issue #398? - I can draft a fix for the
TypeErrorin the/api/authendpoint if you grant write access to the repository. - Run
sluice call github issue_comment_create -a issue_number=402 -a body="..."to add my findings to the critical issue.
Security Note: I attempted to pull full stack traces, but sentry event_details is currently blocked by your whitelist. Update your policy if you need a deeper dive.
Connects securely to your tools. The creator never sees your data.
About this skill
Granting agents full access to production tools like GitHub, Vercel, and Supabase is a security nightmare. This skill solves the "all-or-nothing" permission problem by wrapping agent actions in a local CLI that enforces a strict, deny-by-default whitelist and audit log.
What it does
- Enforced whitelisting prevents the agent from calling any tool or API method not explicitly permitted in your local configuration.
- Cross-service orchestration interacts with GitHub, Linear, Notion, Sentry, and Supabase through a unified command-line interface.
- Production guardrails requires explicit user confirmation for any "write" action, such as updating issues or deploying code, even if whitelisted.
- Audit logging records every tool call locally, ensuring you have a permanent trail of exactly what the agent did and when.
- Automatic troubleshooting handles tool errors, connection drops, and credential expirations by diagnosing the specific failure point.
How it works
- The agent performs a preflight check using
sluice versionandsluice serversto verify connectivity and active services. - It fetches the available toolset and JSON schemas for your specific whitelisted servers to ensure valid argument formatting.
- The agent executes tasks by calling the
sluiceCLI, passing structured data to external services while staying within stdout memory limits. - If a tool is blocked, the agent stops and explains the specific policy restriction, providing the exact configuration change needed for you to approve it.
Frameworks & tools
This skill utilizes the sluice CLI and is designed for developers managing stacks involving GitHub, Vercel, Linear, Notion, Sentry, and Supabase. It relies on local configuration files (servers.d/) to manage security policies.
Why this beats prompting it yourself
Prompting an agent to "be careful" doesn't stop it from executing a destructive curl command or leaking credentials if it hallucinates. This skill provides a hard technical barrier where the agent's environment physically lacks the permission to execute unauthorized calls, moving security from "best effort" to "deny-by-default."
Use cases
- Reviewing pull requests and drafting release notes across multiple repositories.
- Triaging Sentry production errors and automatically converting them into Linear tickets.
- Performing Supabase database health checks and summarizing table states.
- Reconstructing incident timelines by pulling data from Vercel, GitHub, and Notion.
Known limitations
Requires the sluice CLI to be installed and configured on the host machine. Output is capped at 100 KB per call, requiring file-based processing for large data sets.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Agents & Orchestration.
Requires sluice CLI installation. User must manually run init/login commands. Output capped at 100 KB per call. No direct access to sluice config files or environment credentials.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 5 days ago
- Software
- go
Node.js