ClawHub
Back to skill

Security audit

Odds Converter

Security checks for vulnerabilities and agentic risk

Overview

The skill has a clear odds-conversion purpose, but its command examples can turn a malicious odds value into local Python code execution.

Review before installing. The skill appears intended for odds conversion and does not show deception or persistence, but it should be rewritten so user inputs are passed as data, parsed as numbers, and validated before any Python code runs.

A.I.G

Vulnerability Patterns
  • 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
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
Findings (1)

T09 · Insecure Skill Coding Practices

Error
Location
SKILL.md:29
Finding

Arbitrary Python Code Execution Through Unsanitized Odds Substitution

Content
View full analysis
0: impl = 100 / (odds + 100) else: impl = 0.5 dec = round(1 / impl, 4) if impl > 0 else 0 from fractions import Fraction frac = Fraction(1 - impl, impl).limit_denominator(100) if impl > 0 else 'N/A' kalshi = round(impl, 2) print(f'American: {odds:+d}') print(f'Decimal: {dec}') print(f'Fractional: {frac}') print(f'Implied Prob: {impl*100:.2f}%') print(f'Kalshi Price: \${kalshi}') " ``` The same unsafe source-code substitution pattern appears in the other operations: ```bash python3 -c " dec = DEC impl = 1 / dec if dec > 0 else 0 ... " ``` ```bash python3 -c " num, den = NUM, DEN dec = round((num / den) + 1, 4) ... " ``` ```bash python3 -c " impl = PROB dec = round(1 / impl, 4) if impl > 0 else 0 ... " ``` ```bash python3 -c " odds_list = [-150, +200, -110, +300, -400] ... " ``` The batch-conversion instructions explicitly direct the agent to replace the embedded list with supplied values, exposing the same issue if list elements are not independently parsed as numbers. ### Technical Analysis `python3 -c` evaluates its argument as executable Python source. The placeholders `ODDS`, `DEC`, `NUM`, `DEN`, and `PROB` are positioned directly in Python expressions rather than being received through a data-only interface. If an agent performs the req ...[truncated 2765 chars]
Remediation
View remediation

SkillSpector

Vulnerability Patterns
  • System Prompt LeakageDirect Leakage, Indirect Extraction, Tool-Based Exfiltration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (1)

Direct Prompt Extraction

High
Category
System Prompt Leakage
Confidence
85% confidence
Finding

Skill contains instructions that could directly expose system prompts, internal rules, or hidden instructions to users or external parties.

Content

Scanner excerpt · SKILL.md (reported line 150)May include surrounding context.

"

text

## Output Rules

1. Always show ALL five formats in conversion output
2. American odds must include the +/- sign (e.g., +150, -200)