Back to skill

Security audit

NBA-games

Security checks for vulnerabilities and agentic risk

Overview

This skill looks like an NBA schedule helper, but it installs and runs a changeable Python script from GitHub without integrity checks.

Install only if you trust the GitHub repository and are comfortable with the agent running code downloaded from its current main branch. A safer version would bundle the script with the skill or pin and verify the downloaded file before execution.

Vulnerability Patterns
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • 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
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Findings (1)

T03 · Remote Payload Retrieval and Execution

Error
Location
SKILL.md:11
Finding

Mutable Remote Python Payload Is Downloaded and Executed Without Verification

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, lines 11–20; subsequent execution is documented at lines 43–58
Vulnerability Type: Remote payload retrieval and execution
Risk Level: Critical

Vulnerable Code:

bash
SCRIPT_PATH="$HOME/.openclaw/skills/nba_games/nba-schedule.py"
if [ ! -f "$SCRIPT_PATH" ]; then
  mkdir -p "$(dirname "$SCRIPT_PATH")"
  curl -fsSL \
    https://raw.githubusercontent.com/highdeserthacker/nba-schedule/main/nba-schedule.py \
    -o "$SCRIPT_PATH"
fi

The downloaded file is subsequently executed as follows:

bash
# Next 3 upcoming Golden State Warriors games
python3 $HOME/.openclaw/skills/nba_games/nba-schedule.py --team-id 9

# Warriors games in the next 2 days only
python3 $HOME/.openclaw/skills/nba_games/nba-schedule.py --team-id 9 --days-future 2

# Warriors games in the next 2 days plus yesterday's result
python3 $HOME/.openclaw/skills/nba_games/nba-schedule.py --team-id 9 --days-future 2 --days-past 1

# Lakers: next 5 days
python3 $HOME/.openclaw/skills/nba_games/nba-schedule.py --team-id 13 --days-future 5

Technical Analysis

The installation instructions retrieve a Python program from the mutable main branch of a contributor-controlled GitHub repository. The Skill does not pin the resource to an immutable commit, verify a cryptographic digest or signature, validate the downloaded content, or require a review step before execution.

Consequently, the code actually executed can change after the Skill itself has been audited. An attacker who controls or compromises the repository can replace the expected scheduling utility with arbitrary Python code. A fresh installation would download that modified payload, and an ordinary NBA schedule request would execute it through python3.

The if [ ! -f "$SCRIPT_PATH" ] condition does not mitigate initial compromise. It also causes the unverified payload to remain in persistent Skill storage ...[truncated 1786 chars]

Remediation
View remediation

Remediation Suggestions

  1. Bundle the reviewed nba-schedule.py implementation inside the Skill artifact so the audited code is the code that executes.
  2. If remote retrieval is unavoidable, pin the download URL to a specific immutable commit rather than main.
  3. Publish a trusted SHA-256 digest or digital signature with the Skill and verify it before moving the file into place or executing it.
  4. Fail closed if download, signature, ownership, permission, or integrity verification fails.
  5. Download into a securely created temporary file, validate it, assign restrictive permissions, and only then atomically install it.
  6. Review the exact pinned source and restrict its filesystem and network access through sandboxing where supported.
  7. Prefer a small bundled client that directly queries a documented NBA data endpoint, granting only the network access needed to retrieve schedule information.
  8. Correct the inconsistent team lookup command at SKILL.md:37, which refers to /home/node/python/nba-schedule/nba-schedule.py rather than the documented installed path.
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill instructs the agent to fetch executable Python code directly from a third-party GitHub repository into a local skills directory, with no pinning, integrity verification, review step, or sandboxing. Because this skill's purpose is simple NBA schedule lookup, pulling arbitrary remote code creates an unnecessary supply-chain execution path: if the repository is compromised or changed, the agent could run attacker-controlled code on the host.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The installation block performs a network fetch and writes the response to a local executable path without warning the user or requiring confirmation. This reduces visibility into a sensitive action and increases the chance that users or agents will silently import and later execute untrusted code, especially in automated environments where such side effects may be overlooked.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Low
Category
Not specified by scanner
Confidence
88% confidence
Finding

The skill documentation says the installed script lives at $HOME/.openclaw/skills/nba_games/nba-schedule.py, but the Team ID Lookup section tells users to run /home/node/python/nba-schedule/nba-schedule.py --list instead. This actively conflicts with the documented installation and usage path and can mislead auditors or operators about what code is actually being run.

Content

No source excerpt is available for this finding.