Vibe Clawing

PassAudited by ClawScan on May 10, 2026.

Overview

This is an instruction-only coaching skill about delegating work to agents; it has no code or credentials, but users should apply its autonomy advice carefully for high-impact tasks.

This skill appears benign and instruction-only. Treat its autonomy framework as advice, not permission for an agent to act without limits: require approval for high-impact actions, define escalation triggers, keep audit logs, and avoid storing secrets in the suggested local memory files.

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.

What this means

If applied too broadly, an agent could take actions the user would normally want to approve first.

Why it was flagged

The skill explicitly encourages shifting decisions and actions to agents. This is central to the stated purpose, and the text includes guardrails, but users should not apply it blindly to sensitive or irreversible actions.

Skill content
| Decision | You think | Agent decides (with guardrails) |\n| Action | You do | Agent executes |
Recommendation

Keep human approval for financial, legal, account, deployment, deletion, public-posting, or other high-impact actions unless there are explicit safeguards and rollback plans.

NoteHigh Confidence
ASI10: Rogue Agents
What this means

Unattended workflows can produce unwanted changes if the task scope, escalation rules, and stop conditions are unclear.

Why it was flagged

The skill promotes autonomous or unattended operation as a mindset. The artifacts do not implement background agents or hidden persistence, but the advice should be bounded by user-defined stopping conditions.

Skill content
**The shift:** Work continues without you. Loops run whether you're watching or not.
Recommendation

Define what the agent may do unattended, when it must stop, when it must ask, how results are audited, and how changes can be reversed.

What this means

Private or outdated notes could affect future agent behavior if reused without review.

Why it was flagged

The skill asks for persistent local notes that may influence future delegation decisions. This is disclosed and purpose-aligned, but persistent notes can become stale or include sensitive context if the user adds it.

Skill content
Track your vibe clawing journey in `~/vibe-clawing/memory.md`. Create on first use:
Recommendation

Avoid putting secrets or sensitive personal/business data in these notes, and periodically review or delete stale entries.