papadopouloskyriakos/agentic-chatops/cross-tier-review
Claude Code and agent skill by papadopouloskyriakos/agentic-chatops
- community
Checking that the SKILL.md file is still in the repository.
Cross-tier review protocol — chain of verification for REVIEW REQUEST messages from Claude Code (Tier 2). Includes 5-step checklist and REVIEW_JSON output format.
Install
npx vskill@latest install papadopouloskyriakos/agentic-chatops/cross-tier-reviewSKILL.md
Cross-Tier Review Protocol — Chain of Verification
When you see "REVIEW REQUEST:" in an infra room, Claude Code (Tier 2) is asking you to review its work because its confidence was below 0.7. You MUST perform a structured verification before giving your verdict.
Verification Checklist (execute ALL steps)
- CLAIM CHECK: What factual claims does the analysis make? Use the
exectool to
independently verify at least ONE claim (e.g., check a service status, query YT issue, look up a host in NetBox). Do NOT just read and agree.
- ASSUMPTION CHECK: What assumptions are implicit? Flag any that are unsupported
by evidence in the analysis.
- ALTERNATIVE CHECK: What alternative root causes were NOT considered? Name at
least one plausible alternative, even if you think Claude's diagnosis is correct.
- RISK CHECK: Could the proposed action cause secondary failures? Check for
dependencies (e.g., will restarting X affect Y? Does draining node Z displace SeaweedFS pods?).
- RECURRENCE CHECK: If this host/alert has a knowledge base entry, does the proposed
fix address the root cause or just patch the symptom again?
Verdict
After verification, reply with ONE of:
- REVIEW: AGREE — brief reason why the analysis looks correct after verification
- REVIEW: DISAGREE — specific issue found (wrong root cause, unsafe action, missing evidence)
- REVIEW: AUGMENT — additional context to add (things Claude missed, alternative causes)
After your verdict line, output a structured JSON block so the gateway can parse it:
Rules
- ALWAYS verify at least one claim via exec before giving your verdict
- If confidence was < 0.5, pay extra attention to hallucinated fixes
- If you find the proposed action could cause secondary failures, ALWAYS DISAGREE
- Do NOT escalate review requests. This is YOUR job as independent critic.
Dev Task Reviews (CUBEOS/MESHSAT)
For development review requests, adjust the checklist:
- CLAIM CHECK: Verify at least one code claim via codegraph-lookup
- TEST CHECK: Were tests run? Did they pass?
- SCOPE CHECK: Did the changes stay within the issue scope?
- RISK CHECK: Could changes break other modules?
- CONVENTION CHECK: Does the code follow project conventions?
Security
- Security scanPassed, 100 of 100Pattern checks on the skill's files.
- Model reviewNot runA model review is scheduled for the next rescan.
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingopenclaw/skills/cross-tier-review/SKILL.md
Popularity
- Repository stars
- 107
- Repository forks
- 26
- 7-day trend score
- 23
Stars and forks are for the whole GitHub repository, not this skill alone. Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/papadopouloskyriakos/agentic-chatops/cross-tier-review)