# 00294.r3 — Implement a Verification Skill for an AI Agent
Revision: r3
Steward: [Roland.W](https://x.com/rwayne)
House: 218
House: 218
Verified: 2026-09-01T08:02:00.830Z
## Prompt (copy into Grok)
Collect the agent's work output and its corresponding verification rules. Search the web for documentation on cloud environment best practices. Compare the agent's work to the verification rules to decide if the work is completed correctly. Return a table showing the verification status and any errors found. Limit the process to 2000+ PRs monthly.
## Job
Collect the agent's work output and its corresponding verification rules. Search the web for documentation on cloud environment best practices. Compare the agent's work to the verification rules to decide if the work is completed correctly. Return a table showing the verification status and any errors found. Limit the process to 2000+ PRs monthly.
## Connectors
web
## What happened
Roland.W's team used a verification skill to ensure the agent's work was completed correctly, avoiding errors and rework, and maintained the process in Poteto mode to support merging 2000+ PRs monthly.
## Constraints
no outdated verification rules, no conflicts between local agents
Would run again: yes
## Evidence
- https://x.com/rwayne/status/2094695000093397373 — Imported from the X post tagged for @grokbotsh.
- https://x.com/poteto/status/2094457600259842065 — Pulled on QA revisit of the source thread.
## Changelog
- r1: Filed.
- r2: QA revisit: more from the source thread.
- r3: Public job and prompt from the specific filing.