Back to skill
Skillv0.1.0
ClawScan security
Backtest Expert - Strategy Validation · ClawHub's context-aware review of the artifact, metadata, and declared behavior.
Scanner verdict
BenignMay 1, 2026, 7:59 AM
- Verdict
- Benign
- Confidence
- medium
- Model
- gpt-5.5
- Summary
- The provided artifacts describe a coherent backtesting methodology skill with optional local scoring scripts and no evidenced credential use, network access, broker connection, persistence, or trade execution.
- Guidance
- This appears safe from an agentic-security perspective. Before relying on it financially, remember it is a methodology and scoring aid, not a guarantee of profitable trading; review any optional local script before running it and use independent judgment before live deployment.
Review Dimensions
- Purpose & Capability
- okThe README and SKILL.md consistently describe guidance for validating trading strategies and explicitly say the skill does not provide signals, connect to brokers, or execute trades.
- Instruction Scope
- okThe instructions stay within backtesting methodology, robustness checks, bias prevention, and interpretation of results; no prompt override, hidden authority, or autonomous high-impact action is evidenced.
- Install Mechanism
- okThere is no install spec, no required binaries, and no package installation. The included Python files appear to be optional local evaluation/test scripts aligned with the skill purpose.
- Credentials
- okNo environment variables, credentials, API keys, external services, or OS-specific permissions are requested.
- Persistence & Privilege
- okNo background worker, persistent memory, account access, privilege escalation, or long-running autonomous behavior is evidenced in the provided artifacts.