ClawHub

Superpowers Dispatching Parallel Agents

v1.0.0

Dispatch independent tasks to focused agents working concurrently on isolated problems without shared state or sequential dependencies for faster resolution.

⭐ 0· 107·0 current·0 all-time
by@demo112
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Benign
high confidence
✓
Purpose & Capability
Name and description match the SKILL.md content: the document describes how to spawn focused subagents to tackle independent test/code problems in parallel. There are no declarations (env, binaries, or installs) that are out of line with a coordination/prompting pattern.
ℹ
Instruction Scope
Instructions tell agents to read specific test files, identify root causes, and make focused fixes — this is appropriate for a developer troubleshooting pattern. Note: the skill explicitly instructs agents to read and change project files (tests/production code) when tasked, so users should be aware agents following this may modify repository contents; the SKILL.md includes constraints (e.g., 'Do NOT change production code') which help limit scope.
✓
Install Mechanism
No install spec and no code files — lowest-risk model: nothing is written to disk or fetched at install time.
✓
Credentials
No environment variables, credentials, or config paths are requested. The skill does not ask for secrets or unrelated service tokens.
✓
Persistence & Privilege
always:false and default agent invocation are used. This is standard for skills; it does not request elevated or persistent platform privileges and does not modify other skills or global configuration.
Assessment
This skill is an instruction-only pattern for coordinating parallel developer agents and is internally consistent. Before using it: limit the agent's repository scope and avoid giving it unrelated secrets; on first runs, require human review of proposed changes (the SKILL.md already recommends reviewing summaries and running the full test suite); apply it only to truly independent problems (shared-state bugs should be handled sequentially); and consider running agent edits in a branch or sandbox so you can review/CI-test changes before merging.

Like a lobster shell, security has layers — review code before you run it.

latestvk970pvq6e57nwrtey08fryjbds83kkjr

License

MIT-0
Free to use, modify, and redistribute. No attribution required.
Termshttps://spdx.org/licenses/MIT-0.html

Comments