io.github.Kisyntra/agent-sudo-mcp

Add authorization, delegation, provenance, and audit controls to AI agents.

Source

Official Registry

Updated

2026-06-12

// securityCaution

Runs locally
Official / curated
Community-trusted

Summary

The material is sparse, but the tool is confirmed to execute code and its name/description suggest a high-privilege control-plane role for agent authorization, delegation, and audit. Official registry listing and recent updates reduce some source concerns, but lack of source, license, and README limits auditability; overall this warrants caution, with elevated attention on code execution and data access.

CredentialsLow risk

The material states that no keys or environment variables are required, and it does not request API tokens, account passwords, or other explicit credentials; based on the available facts, direct credential exposure appears limited.

Network egressLow risk

The material lists no remote endpoint host and does not declare any external service connections or third-party data transmission; based on the disclosed facts, there is no explicit network egress path.

Code executionCaution

The system flags this tool as executes-code, meaning it can run code or spawn processes locally; combined with the 'authorization, delegation' description, it may participate in higher-privilege agent operations. Code execution is an inherent capability for this class of MCP tools and does not alone justify a high-risk rating, but its runtime privileges and callable scope should be constrained.

Data accessCaution

The description mentions provenance and audit, which commonly implies handling activity logs, context, or local state; however, the material does not specify which files, directories, or system resources it can read or write. No explicit over-privilege red flag is shown, but the permission boundary is opaque and its real data-access scope should be reviewed.

Supply chainCaution

Positive factors include its presence in the official registry and updates within the last year; however, there is no open-source repository, no license, no README, and community adoption is 0 stars, leaving source and dependency behavior unauditable and documentation minimal. Source credibility lowers the risk somewhat, but this still warrants supply-chain caution due to limited auditability.

Recommendations

Install and test it only in an isolated environment, with constrained command execution and system privileges.
Verify its actual readable/writable paths, log contents, and where audit data is stored before enabling it.
Run it under a least-privilege account if possible, and enable process and file-access monitoring.
Prefer to wait for fuller documentation, license terms, or source disclosure before using it in production or with sensitive data.

Audited by: gpt-5.4 · 2026-06-17

// install

Copy the install command and let the AI configure it · recommended for beginners

Please install the "io.github.Kisyntra/agent-sudo-mcp" MCP server from askskill:
Run: claude mcp add 'io-github-kisyntra-agent-sudo-mcp' -- npx -y agent-sudo-mcp