Web3 Target Team Research
AdvisoryAudited by Static analysis on Apr 30, 2026.
Overview
No suspicious patterns detected.
Findings (0)
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.
The agent could keep launching lead-hunting subagents and consuming resources or changing CSV files even after the user thinks the task is finished.
This instructs a scheduled job to silently create new subagents, which can keep autonomous activity running in the background without normal user visibility.
If fewer than 3 running, IMMEDIATELY spawn 3 new hunters using sessions_spawn ... Reply NO_REPLY after checking/spawning.
Do not enable the cron or auto-respawn workflow unless you explicitly want continuous background operation; require visible status updates and a clear user-owned stop command.
A future agent may treat the skill’s old instructions as higher priority than the current user’s intended task boundaries.
The workflow changes the agent’s stopping conditions and ties stopping to a named person rather than the current installing user.
Retry on errors — browser timeouts happen, just retry. If anything times out, RESTART immediately and keep going. Never stop. ... RUN 24/7 UNTIL DERRICK SAYS STOP.
Replace these instructions with bounded run limits, explicit current-user approval, and a normal stop condition for each invocation.
The user may get repeated automated tool use, web browsing, screenshots, and file updates beyond the scope of a single request.
The documented cron job uses agent-session tools on a schedule to spawn multiple subagents without per-run confirmation.
cron({ action: "add", job: { text: "... Use sessions_list ... IMMEDIATELY spawn 3 new hunters using sessions_spawn ...", schedule: "*/10 * * * *", enabled: true } })Require manual confirmation before spawning subagents, cap the number and runtime of hunters, and avoid silent scheduled tool use.
The workspace may continue carrying instructions that influence later agent behavior, even when the user is working on unrelated tasks.
Writing persistent operational instructions into HEARTBEAT.md can cause future sessions to inherit and over-trust the auto-hunting behavior.
Add to your workspace HEARTBEAT.md: ... Keep the search running until told to stop
Keep persistent instructions opt-in, clearly labeled, removable, and scoped to this skill; remove the HEARTBEAT.md entry when the hunt is done.
Users have limited ability to confirm who authored the workflow or whether the persistent automation guidance is appropriate.
There is no executable package here, but the skill’s provenance is not verifiable from the supplied metadata.
Source: unknown; Homepage: none; No install spec — this is an instruction-only skill.
Prefer skills with a clear source or repository, especially before enabling cron jobs or persistent agent instructions.