Track bankroll across sportsbooks and prediction markets. Log bets, record results, calculate ROI, generate P&L reports, and enforce risk limits. Use when as...

This skill mostly looks like a simple local bankroll tracker, but take these precautions before installing or invoking it: - Verify the python3 requirement: ask the author or inspect additional documentation. If the skill never uses Python, the metadata may be inaccurate. - Backup and inspect ~/.openclaw/data/bankroll.db (and create the DB schema yourself) before letting the agent write to it. The SKILL.md provides only an INSERT example and no schema; running commands as-is could corrupt data. - Confirm file permissions for ~/.openclaw/data so only you/your agent can access it — the skill will read/write a file in your home dir. - Follow the AgentBets links to verify the publisher and documentation; the skill registry shows no homepage/source, which reduces provenance confidence. - Because this skill executes sqlite3 commands, ensure the running agent environment actually contains sqlite3 (and that you trust it to run local commands). If you want less risk, require manual review of any DB-modifying commands before they run. If you cannot verify the missing pieces (schema, why python3 is needed, or the publisher identity), treat the skill cautiously or prefer a vetted alternative.

Type: OpenClaw Skill Name: bankroll-manager Version: 1.1.0 The bankroll-manager skill provides templates for executing raw sqlite3 commands via the shell to manage a local database (~/.openclaw/data/bankroll.db). This design is highly vulnerable to SQL injection and potential shell injection, as it instructs the agent to interpolate user-provided strings (e.g., 'NOTES', 'SELECTION') directly into command-line arguments. While no evidence of intentional malice or data exfiltration was found in SKILL.md, the reliance on unparameterized shell-based database manipulation is a significant security risk.