Market Watch

PassAudited by ClawScan on May 1, 2026.

Overview

Market Watch is a coherent market-alert skill, but users should notice that it runs background monitors, stores alert/session context locally, and can proactively deliver messages to the agent.

Install this only if you are comfortable with a background market/news monitor that stores alert context locally and can proactively message your agent. Review recipients and alert summaries, avoid storing secrets, and stop or uninstall the daemon/watchdog when monitoring is no longer needed.

Findings (4)

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

What this means

After an alert is registered, market/news monitoring can continue in the background and may be restarted automatically if the optional watchdog is installed.

Why it was flagged

The skill intentionally uses background daemons and offers an auto-restart watchdog so alerts can fire later.

Skill content
Both scripts automatically start the daemon after registering an alert. ... (macOS) Install launchd watchdog ... Automatically resurrects the daemon every 5 minutes if it crashed
Recommendation

Only install the watchdog if you want persistent monitoring; use the provided cancel, stop, and uninstall commands when you no longer need alerts.

What this means

Local alert files may retain sensitive trading intentions or conversation references until cancelled or deleted.

Why it was flagged

Alert state persists locally and can include financial context, session identifiers, reply routing, and a transcript path used when an alert fires.

Skill content
所有警报存入 `~/.openclaw/agents/{agent}/private/market-alerts.json` ... `context_summary`, `session_key`, `reply_channel`, `reply_to`, `transcript_file`
Recommendation

Avoid putting secrets in alert summaries, keep the agent private directory protected, and periodically list/cancel old alerts.

What this means

A triggered alert can cause the agent to proactively message a configured channel or session.

Why it was flagged

The monitor sends alert messages into OpenClaw sessions using routing values stored in each alert.

Skill content
cmd = ["openclaw", "agent", "--deliver", "--reply-account", agent_id, "--reply-channel", reply_channel] ... subprocess.Popen(cmd
Recommendation

Check alert recipients and session keys when registering alerts, and treat triggered market/news content as information to verify before acting.

What this means

Dependency versions may vary between installations, which can affect reliability or supply-chain exposure.

Why it was flagged

The setup instructions rely on manually installing unpinned Python packages.

Skill content
pip3 install requests pytdx
Recommendation

Install dependencies from trusted package indexes, consider pinning versions in your environment, and review updates before upgrading.