Skills/PyPI/universal-mcp-agents

universal-mcp-agents@0.1.32

VERIFIED· install only

Publisher claims: "Add your description here" Observed in sandbox: no network, touches nothing outside the project, writes no memory.

Registry PyPI (MCP server)Publisher universal-mcp · publishing for 556 daysInstalls 259Last detonated 21 min ago
$claude mcp add universal-mcp-agents -- uvx universal-mcp-agents
WatchReport an issue

Coverage

Bundled scripts ran with no arguments; no agent task scenario for this bundle.
Tested
8 of 51 bundled scripts run with no arguments (the rest were not run: per-bundle cap), with decoy credentials in the environment · every file read, file written and host contacted by what ran
Not tested
use of the skill by an agent

What this means for the label: Verified here covers what ran. The server's tools were not exercised with a real credential, so nothing about their behavior is asserted; the credential summary below says what to do with that.

Before you give it your token

It will ask for
AGENTR_API_KEYAWS_ACCESS_KEY_IDAWS_SECRET_ACCESS_KEYRESEND_API_KEYConfiguration, not credentials: AGENTR_BASE_URL, AWS_REGION, CHAT_MODEL, FRONTEND_URL, LLM_TOOLS_LLM
Where it talks to
no vendor host named in the codeAlso ruzo.ai, agentr.dev, tmpfiles.org, tailwindcss.com, wikimedia.org — not the publisher's domain; check what for before connecting.
What we saw when it ran
no credential files read · no network
What we could not test
use of the skill by an agent

Connect it with a scoped credential, under the guard, and read the third-party hosts first. The code names ruzo.ai, agentr.dev, tmpfiles.org, tailwindcss.com, wikimedia.org as well as its own vendor. The guard injects AGENTR_API_KEY, AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY, RESEND_API_KEY only into requests to the host you bind (and its subdomains) and refuses the rest.

$vet-skills run --net declared -- uvx universal-mcp-agents

Never paste a credential into a chat, a ticket or a report, including ours. How the guard holds it.

Is [email protected] safe to install?

KillerSkills ran it in the KillerSkills honeypot sandbox (2 scenarios; bundled scripts ran with no arguments; no agent task scenario for this bundle) on 2026-11-20 and scored it 80/100 (Verified). In what ran, nothing we could detect went wrong: no undeclared network egress, no credential reads, no persistent instruction writes. What it does with a real credential was not observed; the static findings and the publisher carry the rest of the score. Summary: no network, touches nothing outside the project, writes no memory.

Behavior trace

Honeypot run · 1 s · 2 scenarios · 0 deviations
t+CategoryObserved action
0.0sprocessno install scripts in bundle
0.0sprocessbundled script run in the sandbox with no arguments
0.1sprocessexited 1: ModuleNotFoundError: No module named 'dotenv'
0.1sprocessbundled script run in the sandbox with no arguments
0.1sprocessbundled script run in the sandbox with no arguments
0.2sprocessexited 1: ModuleNotFoundError: No module named 'openevals'
0.2sprocessbundled script run in the sandbox with no arguments
0.2sprocessbundled script run in the sandbox with no arguments
0.3sprocessexited 1: ModuleNotFoundError: No module named 'typer'
0.3sprocessbundled script run in the sandbox with no arguments · not named in the instructions
0.4sprocessexited 1: ModuleNotFoundError: No module named 'universal_mcp'
0.4sprocessbundled script run in the sandbox with no arguments · not named in the instructions
0.4sprocessexited 1: ModuleNotFoundError: No module named 'typer'
0.4sprocessbundled script run in the sandbox with no arguments · not named in the instructions
0.5sprocessexited 1: ImportError: cannot import name 'Any' from partially initialized module 'typing' (most likely due to a circular import) (/usr/lib/python3.11/typing.py)
0.5sprocess8 of 51 bundled scripts run with no arguments (the rest were not run: per-bundle cap)
0.5sprocessscript backend runs setup only
0.5sprocessscript backend runs setup only
0.5sprocessno install scripts to re-run

Static findings

HIGH
Access to sensitive file pathsCLAUDE.md:198 · Access to sensitive file paths — SSH keys, AWS credentials, env files, or system files- Requires `.env` file with auth tokensFix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.
AST05
MEDIUM
Python exec or eval of computed codesrc/universal_mcp/agents/Sandbox/sandbox_worker.py:148 · Python exec or eval of computed codecoroutine = eval(compiled_code, namespace, namespace)Fix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.
AST05
MEDIUM
File type combined with upload actionsrc/evals/search.py:7 · File type combined with upload action — agent may exfiltrate local data files`src/evals/datasets/search.json`, executes searches, computes simple accuracyFix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.
AST07
HIGH
Access to sensitive file pathsGEMINI.md:198 · Access to sensitive file paths — SSH keys, AWS credentials, env files, or system files- Requires `.env` file with auth tokensFix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.
AST05
HIGH
Access to sensitive file pathsAGENTS.md:198 · Access to sensitive file paths — SSH keys, AWS credentials, env files, or system files- Requires `.env` file with auth tokensFix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.
AST05
HIGH
Access to sensitive file pathslanggraph.json:7 · Access to sensitive file paths — SSH keys, AWS credentials, env files, or system files"env": ".env",Fix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.
AST05
LOW
Prompt injectionsrc/universal_mcp/agents/codeact0/todos.md:16 · Prompt injection — known jailbreak technique reference · informational: not an instruction file- [ ] Setup jailbreak guardrailsFix: Keep SKILL.md to visible, descriptive text. Remove comments and encoded text that address the agent.
AST01
MEDIUM
Python exec or eval of computed codesrc/universal_mcp/agents/Sandbox/in_process_sandbox.py:184 · Python exec or eval of computed codecoroutine = eval(compiled_code, self.namespace, self.namespace)Fix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.
AST05
MEDIUM
Python exec or eval of computed codesrc/universal_mcp/agents/Sandbox/sandbox_worker.py:246 · Python exec or eval of computed codeexec(class_def, namespace)Fix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.
AST05
MEDIUM
Python exec or eval of computed codesrc/universal_mcp/agents/Sandbox/sandbox_worker.py:255 · Python exec or eval of computed codeexec(func_def, namespace)Fix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.
AST05
MEDIUM
Python exec or eval of computed codesrc/universal_mcp/agents/Sandbox/sandbox_worker.py:237 · Python exec or eval of computed codeexec(import_stmt, namespace)Fix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.
AST05
MEDIUM
Python subprocess execution detectedsrc/universal_mcp/agents/Sandbox/subprocess_sandbox.py:187 · Python subprocess execution detectedprocess = subprocess.Popen(Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
AST05
MEDIUM
Python exec or eval of computed codesrc/universal_mcp/agents/Sandbox/sandbox_worker.py:231 · Python exec or eval of computed codeexec("File = str", namespace)Fix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.
AST05

Cost to the agent

Instructions loaded
≈7,013 tok
4 files · 81 in bundle

Tokens estimated at four characters each. Skills load their instruction text when triggered; the figure is what that costs.

Known vulnerabilities

Not checked against OSV yet; the weekly check has not reached this version.