Agent Anchor
ReviewAudited by ClawScan on May 10, 2026.
Overview
This skill needs review because it relies on an unspecified external dashboard while promising to continuously retain your agent’s state with little detail on data scope or retention.
Before installing, verify the actual repository and inspect the dashboard code. Confirm exactly what agent state is saved, whether sensitive chats/files/credentials can be excluded, where state-anchor.json or other state files live, how to delete them, and whether any cron-job fixes require explicit review before changes are made.
Findings (3)
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.
Private agent context, task history, contacts, revenue information, or cron-job details could be retained longer or more broadly than the user expects.
The skill explicitly advertises persistent agent memory and recurring snapshots, including potentially sensitive operational and business data, but does not document storage location, retention, exclusions, or access controls.
Your AI agent's memory that survives everything... Automatically saves state every 3 minutes... Edit `state-anchor.json` to customize: ... Revenue streams ... Contacts ... Cron jobs to monitor
Use only after inspecting the actual implementation and confirming what is saved, where it is stored, how to delete it, and how sensitive data can be excluded.
The reviewed artifact does not show what code the user would actually run or what that code would do with saved agent state.
The installation instructions require cloning an unspecified external repository, while the provided package contains no code files or install spec. This is material because that unreviewed repository would implement the dashboard that handles persistent agent state.
git clone <agent-anchor-repo> cd agent-anchor python3 -m http.server 3456
Do not install from a placeholder or unknown repository. Verify the repository URL, source owner, commit/version, and code behavior before running the dashboard.
If implemented broadly, one-click fixes could change local scheduled tasks in ways the user may not fully review.
The skill advertises the ability to fix cron-job errors, which could imply changing scheduled jobs or local automation. The artifacts do not describe the scope or reversibility of those fixes.
Cron Job Monitoring — Monitor scheduled jobs, click to see details, fix errors with one click
Require a preview and confirmation of any cron-job changes, and verify that fixes are limited to the intended jobs.