Writes ralph loops for you that you can copy and paste

WarnAudited by ClawScan on May 10, 2026.

Overview

This skill only writes commands, but the commands it recommends can run AI coding assistants repeatedly with safety confirmations disabled.

Review the generated shell command before pasting it. Prefer fixed iteration limits and stop conditions, remove no-approval flags unless you fully trust the prompt and working directory, and keep secrets out of PROMPT.md.

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.

High

#ASI02: Tool Misuse and Exploitation

What this means

If the user copies and runs the generated command, the AI CLI could modify files, run tools, or consume account resources repeatedly without normal confirmation prompts.

Why it was flagged

The generated default commands disable or bypass permission prompts for AI coding CLIs, making high-impact tool actions easier to run without per-action review.

Skill content

Default: `Get-Content PROMPT.md -Raw | claude-code --dangerously-skip-permissions` ... Default: `Get-Content PROMPT.md -Raw | gemini --yolo` ... `grok-auto` ... auto-approves permissions

Recommendation

Do not make no-approval flags the default. Require an explicit opt-in warning, prefer safer approval modes, and show users how to remove `--dangerously-skip-permissions`, `--yolo`, or auto-approve behavior.

Medium

#ASI08: Cascading Failures

What this means

A bad prompt, wrong working directory, or mistaken generated command could repeat many times, causing escalating file changes, costs, or workflow corruption.

Why it was flagged

The skill offers loop patterns that can continue indefinitely, which can amplify mistakes when combined with automated AI CLI execution.

Skill content

"Infinite with delay" - "Run forever with pauses"

Recommendation

Make bounded loops the default, require explicit confirmation for infinite loops, include visible stop instructions, and recommend max-iteration or time-limit safeguards.

Low

#ASI07: Insecure Inter-Agent Communication

What this means

Anything placed in PROMPT.md may be processed by the selected external AI service when the command is run.

Why it was flagged

The generated commands intentionally send the contents of PROMPT.md to the selected AI CLI/provider. This is aligned with the skill purpose, but users should understand the data flow.

Skill content

`cat PROMPT.md | claude-code --dangerously-skip-permissions` and `Get-Content PROMPT.md -Raw | gemini --model ... --yolo`

Recommendation

Review PROMPT.md before running generated commands and avoid including secrets, private data, or files you do not want sent to the chosen provider.