Back to skill

Security audit

yuketang-mcp

Security checks for vulnerabilities and agentic risk

Overview

The skill mostly fits a Rain Classroom assistant, but its installers handle credentials unsafely, run unpinned external code, and one script silently reports install telemetry.

Review this skill before installing. Use it only if you trust the YuKeTang/Rain Classroom MCP service and are comfortable granting it access to your Rain Classroom account data and class scheduling actions. Treat YUKETANG_SECRET like a password, avoid pasting it into chats or logs, and prefer a safer pinned setup path over the included unpinned npx installers; the shell installer also sends an undisclosed install report.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
Findings (3)

T09 · Insecure Skill Coding Practices

Error
Location
setup.js:13
Finding

Shell Command Injection Through YUKETANG_SECRET

Content
View full analysis
Remediation
View remediation

T08 · Insecure Dependencies

Warning
Location
setup.sh:44
Finding

Execution of an Unpinned npm Package Through npx

Content
View full analysis
/dev/null 2>&1; then echo "🧪 验证配置..." if npx mcporter list 2>&1 | grep -q "yuketang-mcp"; then echo "✅ 验证通过" npx mcporter list | grep -A 1 "yuketang-mcp" || true else echo "⚠️ 验证未通过,但配置可能已写入" echo " 你可以直接尝试使用,如果失败请检查 Secret 是否正确" fi fi # ── 4. 上报(静默) ── END=$(now_ms) DURATION=$((END - START)) if command -v npx >/dev/null 2>&1; then npx mcporter call yuketang-mcp claw_report \ --args "{\"payload\":{\"durationMs\":${DURATION}},\"action\":\"install\"}" \ >/dev/null 2>&1 || true fi ``` From `setup.js`: ```javascript try { execSync( `npx mcporter config add yuketang-mcp --url "${MCP_URL}" --header "Authorization=${authorization}" --scope project`, { stdio: "inherit" } ); console.log("✅ 注册成功\n"); registered = true; } catch { console.log("⚠️ mcporter 注册失败,请手动配置 ↓\n"); } ``` ```javascript const list = execSync("npx mcporter list", { encoding: "utf8" }); ``` From `SKILL.md`: ```bash npx mcporter call yuketang-mcp ykt_classroom_statistics --args '{"classroomName": "xxx"}' ``` ### Technical Analysis The project repeatedly invokes `npx mcporter` without declaring `mcporter` as a locked project dependency and without specifying an exact reviewed version. The provided `package.json` contains only OpenClaw metadata and has no dependency declaration or lockfile. Depending on the local npm/npx version and environment, `npx` can resolve and download a package that is not already instal ...[truncated 1419 chars]
Remediation
View remediation

other

Note
Location
setup.sh:103
Finding

Undisclosed Authenticated Installation Telemetry

Content
View full analysis
/dev/null 2>&1; then npx mcporter call yuketang-mcp claw_report \ --args "{\"payload\":{\"durationMs\":${DURATION}},\"action\":\"install\"}" \ >/dev/null 2>&1 || true fi ``` ### Technical Analysis After configuration, the shell installer invokes the remote `claw_report` MCP tool and transmits an `install` action together with the setup duration. The call uses the previously registered YuKeTang MCP service, which was configured with the user's bearer credential. The operation is explicitly marked as silent, and both standard output and standard error are discarded. The user-facing installation documentation does not disclose this telemetry call or provide a consent or opt-out mechanism. Although the visible payload is limited to an action and duration, the authenticated remote service can associate the event with the account represented by `YUKETANG_SECRET` and may also receive ordinary connection metadata. ### Attack Path 1. The user exports a valid `YUKETANG_SECRET`. 2. The user runs `bash setup.sh`. 3. The script registers the authenticated remote MCP endpoint. 4. After setup, it silently invokes `claw_report` with `action` set to `install` and the measured installation duration. 5. Output and errors are suppressed, so the user receives no indication that the report was transmitted or attempted. ### Impact Assessment The remote service can learn that the authenticated user installed the Skill and can receive timing and standard network metadata. This is a privacy and transparency issue rather than a direct privilege escalation. No evidence in the reviewed code shows transmission of local file contents or the raw secret in the telemetry payload. However, the cal ...[truncated 90 chars]
Remediation
View remediation
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
Findings (21)

Tp4

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The declared purpose presents the skill as a data-query tool, but the content also instructs local secret management, environment-variable updates, installation steps, and external command execution. This mismatch reduces informed consent and can trick users or reviewers into underestimating the skill's ability to influence local configuration and credential handling.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
93% confidence
Finding

The skill clearly relies on sensitive local capabilities via environment variables (YUKETANG_SECRET), but it does not declare any explicit tool scope or permission boundary. That makes the skill harder to govern and review, and can lead to unintended credential access or broader execution than users expect from a 'query' skill.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
91% confidence
Finding

Using npx mcporter without pinning a specific version creates a supply-chain risk because the executed package may change over time or resolve to a malicious/compromised release. Since this skill instructs installation and runtime use of the tool, an attacker controlling the package path or a bad upstream update could gain code execution on the user's machine.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The skill tells users to obtain a personal Secret and even says the agent should help update environment variables, but it does not warn that the Secret is a sensitive credential that should not be pasted into chats, logs, or unsafe contexts. This increases the chance of credential leakage through transcripts, shell history, screenshots, or overbroad agent handling.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The instruction says all relative times must use current system Beijing time as the basis, which forces a specific locale/timezone behavior. Because the file does not offer user choice or explain that the skill is limited to a China-specific operating context, this is a natural-language locale policy violation under the stated rules.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
91% confidence
Finding

This is another instance of unpinned npx mcporter usage in operational guidance, carrying the same supply-chain and arbitrary-code-execution risk. Repeated unpinned invocation increases exposure because users may run it during setup and troubleshooting.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
91% confidence
Finding

The example command at this line also executes mcporter through unpinned npx, so users copying the example are exposed to the same supply-chain compromise path. Example code is especially risky because it is likely to be executed verbatim.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

This markdown file contains user-facing instructions, examples, and prompts exclusively in Chinese, including the title and all interaction guidance. Under the policy rule for natural-language violations, forcing a single language without user opt-in can be a locale policy issue unless the regional restriction is explicitly justified.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The reference documentation exposes an action-capable tool (cube_lesson_reservation) for scheduling classes, while the skill metadata describes the skill as providing account and class query services. This mismatch can cause an orchestrator, reviewer, or end user to under-scope the skill's authority and accidentally permit state-changing operations that were not expected, increasing the risk of unintended class creation or meeting scheduling.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The script passes a bearer token on the command line to npx mcporter, which can expose the secret to local process inspection, shell history/logging, or downstream tooling, and it modifies local MCP configuration without a clear warning or consent prompt. In this skill context, the token grants access to a classroom/account data service, so accidental disclosure could expose educational account or class information.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The script instructs the user to talk to the AI using a Chinese phrase, and the surrounding comments and messages are also exclusively in Chinese. This imposes a language expectation without user opt-in or explanation that the skill is intentionally Chinese-only.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
84% confidence
Finding

The script invokes npx mcporter without pinning an exact package version, so installation or execution may resolve to whatever version is currently available from the registry or local environment. In a setup script that handles bearer tokens and writes MCP configuration, this creates a supply-chain risk where a compromised or incompatible package could capture secrets or alter client configuration.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
84% confidence
Finding

This verification step also executes npx mcporter without a pinned version, exposing the user to the same package-substitution or malicious-update risk during post-install validation. Although this step is read-oriented, it still runs external code in the user's environment and could be abused if the package source is compromised.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
84% confidence
Finding

The script again calls npx mcporter without version pinning when printing the configured entry, which keeps the attack surface open even after setup appears complete. Repeated unpinned external execution increases exposure to registry compromise, typosquatting, or unexpected package updates.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The setup script performs an additional claw_report action after configuration, which is unrelated to the user's core task of registering the MCP service. Extra remote actions in installers are risky because they expand the behavior beyond expected local setup and may transmit metadata or trigger server-side processing without clear necessity.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The script silently sends installation telemetry (durationMs and install action) to a remote service with output suppressed and no user-facing warning or consent. Silent telemetry undermines transparency and trust, and in security-sensitive setup flows it can normalize undisclosed data transmission that may later expand to more sensitive fields.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
87% confidence
Finding

The telemetry/reporting path executes npx mcporter without version pinning while contacting the remote MCP service, combining supply-chain risk with networked data transmission. Because the script has access to the bearer token and installation context, a malicious package version could exfiltrate credentials or send unauthorized actions.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Low
Category
Not specified by scanner
Confidence
85% confidence
Finding

The manifest describes a rain-classroom query skill for account, class, warning-list, teaching, and homework-announcement lookups. In this file, the implemented behavior is local environment inspection and subprocess execution via npx mcporter ..., which is a packaging/setup capability rather than part of the declared query service itself.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
88% confidence
Finding

The final instruction tells the user to speak to the AI using a Chinese phrase, and the script's user-facing text is entirely in Chinese without offering an alternative language or opt-in. This can violate language/locale policy when a skill implicitly requires a specific language without giving the user a choice.

Content

No source excerpt is available for this finding.