Type-Based Autonomy
PassAudited by ClawScan on May 10, 2026.
Overview
This instruction-only skill is coherent for type-filtered autonomous task work, but users should understand it can keep working from a local queue and persist notes if they enable the heartbeat workflow.
Before installing, decide whether you really want heartbeat-style autonomous work. Edit the RA/MONEY goal text to your own goals, confirm cron or another system handles skipped backup/security/maintenance tasks, and keep the queue and memory files under review or version control.
Findings (5)
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.
If integrated into a heartbeat process, the agent may spend tokens and make progress on queued tasks without a fresh prompt for each task.
The heartbeat protocol allows the agent to continue selecting and performing queued tasks. This is the skill's disclosed purpose, but it is autonomous behavior users should intentionally enable.
Task complete and time remaining? → Pick another allowed-type task if queue has one
Enable the heartbeat only when desired, keep the queue reviewed, and use the provided token/time limits.
Persistent notes may preserve goals, preferences, file states, and resume instructions across sessions.
The skill tells future sessions to rely on persisted checkpoint content. This supports recovery, but stale or tampered memory files could steer later agent behavior.
Checkpoints include "Resume Instructions" — follow them exactly
Store these memory files only in trusted workspaces, review them periodically, and avoid placing secrets or untrusted instructions in them.
If left unchanged, the agent may bias task selection and follow-up work toward an RA/MONEY objective rather than the user's own goals.
The template contains a specific hard-coded goal and identity reference. It is visible, but may not match every user's intended objective.
Every task should link to RA's long-term goal: **MONEY**
Edit GOALS.md and the queue template to reflect the actual user's goals before enabling autonomous work.
If no separate cron process is actually configured, skipped maintenance, backup, or security tasks may not run.
The autonomy workflow deliberately skips maintenance, backup, and security tasks because it assumes separate cron jobs handle them.
Maintenance (`@type:maintenance`) - Cron handles cleanup, backup ... Security (`@type:security`) - Monthly audit by cron
Confirm the separate cron jobs exist, or change the queue policy so important operational tasks are handled elsewhere.
Users cannot verify the exact upstream commit from the supplied metadata.
The provenance link is a placeholder, and the registry source is unknown. Because there is no code or install script, this is a low-impact provenance note rather than a behavior concern.
"commit": "https://github.com/openclaw/skills/commit/placeholder"
Prefer a version with a real source link or review the included instruction files before use.
