← Back to Skills Marketplace
220
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install verify-before-done-avi
Description
Prevent premature completion claims, repeated same-pattern retries, and weak handoffs. Use this skill to improve verification, strategy switching, and blocke...
Usage Guidance
This skill is a behavioral guideline (instruction-only) and appears internally coherent. It will direct the agent to inspect context, files, logs, configs, and environment state when needed — useful for verification but potentially able to read any data the agent already has access to. Before installing or enabling it, confirm the agent's tool/file access policies and whether you are comfortable allowing the agent to look through those resources. If you want tighter control, restrict the agent's file/tool permissions, or only enable the skill for trusted agents/sessions. Confidence would increase further if you provide information about which tools or file-scopes the agent will be allowed to use when following these instructions.
Capability Analysis
Type: OpenClaw Skill
Name: verify-before-done-avi
Version: 1.0.0
The skill bundle 'verify-before-done' contains meta-instructions designed to improve the AI agent's operational discipline, such as enforcing verification of tasks, avoiding repetitive failed strategies, and providing high-quality handoffs. There is no evidence of malicious code, data exfiltration, or harmful prompt injection in SKILL.md or _meta.json.
Capability Assessment
Purpose & Capability
Name and description match the SKILL.md instructions: the skill is a behavioral guideline for verifying work, avoiding repeated unsuccessful attempts, and producing high-signal handoffs. It does not request binaries, env vars, or installs that would be unrelated to this purpose.
Instruction Scope
SKILL.md instructs agents to check context, files, logs, docs, tool output, environment state, repo structure and configs when applicable. That is consistent with verification purposes, but gives the agent broad discretionary scope to access available files/outputs — sensible for verification but worth being aware of if you restrict agent file or tool access.
Install Mechanism
No install spec or code files; instruction-only skills are lowest install risk and nothing is written to disk or fetched at install time.
Credentials
The skill requests no environment variables or credentials. The SKILL.md advises checking environment state and provided files when necessary — reasonable for verification and not disproportionate, but it implies the agent may read any data it already has access to.
Persistence & Privilege
always:false and normal model invocation settings. The skill does not request persistent presence or modify other skills. Autonomous invocation is allowed (platform default) but not unusual here.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install verify-before-done-avi - After installation, invoke the skill by name or use
/verify-before-done-avi - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
Initial release.
- Introduces rules for verifying completion, avoiding premature claims, and discouraging repeated ineffective retries.
- Provides checklists for switching strategies, performing pragmatic verification, and improving handoff quality when blocked.
- Clarifies response patterns for verified, partially verified, and blocked situations.
- Emphasizes lightweight, evidence-driven checks and appropriate effort scaling.
- Outlines boundaries to ensure the skill remains focused and does not create unnecessary bureaucracy.
Metadata
Frequently Asked Questions
What is verify-before-done?
Prevent premature completion claims, repeated same-pattern retries, and weak handoffs. Use this skill to improve verification, strategy switching, and blocke... It is an AI Agent Skill for Claude Code / OpenClaw, with 220 downloads so far.
How do I install verify-before-done?
Run "/install verify-before-done-avi" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is verify-before-done free?
Yes, verify-before-done is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does verify-before-done support?
verify-before-done is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created verify-before-done?
It is built and maintained by aviclaw (@aviclaw); the current version is v1.0.0.
More Skills