aiskillstore/marketplace / skills/2dmurali/review-loop
review-loop
VerifiedIterative worker-reviewer cycle that spawns a critic subagent to score work 1-10 and provide actionable feedback, then revises until a quality gate is met. Use when implementing features, writing specs, reviewing existing code, or completing any task where quality matters more than speed. Trigger phrases: "use review-loop", "polish this", "iterate on this", "/review-loop", "review with feedback loop".
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 12.2 KB |
| skill-report.json | config | 23.8 KB |
skill-report.json
View on GitHub{
"schema_version": "2.0",
"meta": {
"generated_at": "2026-08-06T01:47:57.641Z",
"slug": "2dmurali-review-loop",
"source_url": "https://github.com/2dmurali/review-loop-skill/tree/0980fb7430997f7c8b26c694c9ff797ce7602378/review-loop",
"source_ref": "0980fb7430997f7c8b26c694c9ff797ce7602378",
"model": "codex",
"analysis_version": "3.0.0",
"source_type": "community",
"content_hash": "0e70af372446c89c7c9dec502b3b1b9a0403c16783e767ae31e1cdb2005fbd29",
"tree_hash": "4f664d5ca72ff965d46d2f0f5571e162a0a981dce9f5bd68fefded3cb5762394",
"provenance": {
"model_effective": "codex",
"fallback_chain": [
{
"agent": "codex",
"outcome": "success"
}
],
"analysis_version": "3.0.0"
},
"upstream_version_raw": null,
"upstream_version_normalized": null,
"upstream_version_source": "none",
"upstream_version_status": "missing",
"upstream_commit_sha": "0980fb7430997f7c8b26c694c9ff797ce7602378"
},
"skill": {
"name": "review-loop",
"description": "Runs iterative worker and reviewer cycles, applies targeted feedback, and stops when a defined quality score or loop limit is reached.",
"summary": "Improve code, documents, and specifications through scored review and targeted revision.",
"icon": "π¦",
"author": "2dmurali",
"license": "MIT",
"category": "productivity",
"tags": [
"iterative-review",
"quality-assurance",
"subagents",
"code-review",
"writing-review"
],
"supported_tools": [
"claude",
"codex",
"claude-code"
],
"risk_factors": [
"external_commands"
]
},
"security_audit": {
"risk_level": "safe",
"is_blocked": false,
"safe_to_publish": true,
"analysis_status": "ok",
"summary": "All 22 static findings are false positives caused by Markdown backticks, fenced examples, or ordinary review guidance. The skill contains no executable scripts, command invocations, reconnaissance instructions, or prompt injection attempts.",
"remediation": [],
"risk_factor_evidence": [
{
"factor": "external_commands",
"evidence": [
{
"file": "SKILL.md",
"line_start": 25,
"line_end": 25
},
{
"file": "SKILL.md",
"line_start": 94,
"line_end": 99
},
{
"file": "SKILL.md",
"line_start": 99,
"line_end": 114
},
{
"file": "SKILL.md",
"line_start": 114,
"line_end": 126
},
{
"file": "SKILL.md",
"line_start": 126,
"line_end": 146
},
{
"file": "SKILL.md",
"line_start": 146,
"line_end": 188
},
{
"file": "SKILL.md",
"line_start": 188,
"line_end": 198
},
{
"file": "SKILL.md",
"line_start": 198,
"line_end": 201
},
{
"file": "SKILL.md",
"line_start": 201,
"line_end": 248
},
{
"file": "SKILL.md",
"line_start": 248,
"line_end": 257
},
{
"file": "SKILL.md",
"line_start": 257,
"line_end": 263
},
{
"file": "SKILL.md",
"line_start": 263,
"line_end": 272
},
{
"file": "SKILL.md",
"line_start": 272,
"line_end": 295
},
{
"file": "SKILL.md",
"line_start": 295,
"line_end": 304
},
{
"file": "SKILL.md",
"line_start": 304,
"line_end": 308
}
]
}
],
"static_findings": [
{
"id": "external_commands:SKILL.md:25:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 25,
"line_end": 25,
"snippet": "1. Say: `\"implement X, use review-loop\"` or `\"run review-loop on the file I just wrote\"`"
},
{
"id": "external_commands:SKILL.md:94:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 94,
"line_end": 99,
"snippet": "```"
},
{
"id": "external_commands:SKILL.md:99:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 99,
"line_end": 114,
"snippet": "```"
},
{
"id": "external_commands:SKILL.md:114:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 114,
"line_end": 126,
"snippet": "```mermaid"
},
{
"id": "external_commands:SKILL.md:126:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 126,
"line_end": 146,
"snippet": "```"
},
{
"id": "external_commands:SKILL.md:146:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 146,
"line_end": 188,
"snippet": "```"
},
{
"id": "external_commands:SKILL.md:188:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 188,
"line_end": 198,
"snippet": "```"
},
{
"id": "external_commands:SKILL.md:198:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 198,
"line_end": 201,
"snippet": "```"
},
{
"id": "external_commands:SKILL.md:201:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 201,
"line_end": 248,
"snippet": "```"
},
{
"id": "external_commands:SKILL.md:248:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 248,
"line_end": 257,
"snippet": "```"
},
{
"id": "external_commands:SKILL.md:257:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 257,
"line_end": 263,
"snippet": "```"
},
{
"id": "external_commands:SKILL.md:263:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 263,
"line_end": 272,
"snippet": "```"
},
{
"id": "external_commands:SKILL.md:272:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 272,
"line_end": 295,
"snippet": "```"
},
{
"id": "external_commands:SKILL.md:295:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 295,
"line_end": 304,
"snippet": "```"
},
{
"id": "external_commands:SKILL.md:304:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 304,
"line_end": 308,
"snippet": "Reviewer: Clean. Minor: variable name `d` could be more descriptive."
},
{
"id": "blocker:SKILL.md:114:system-reconnaissance",
"category": "blocker",
"severity": "low",
"pattern": "System reconnaissance",
"file": "SKILL.md",
"line_start": 114,
"line_end": 115,
"snippet": "```mermaid"
},
{
"id": "blocker:SKILL.md:215:system-reconnaissance",
"category": "blocker",
"severity": "low",
"pattern": "System reconnaissance",
"file": "SKILL.md",
"line_start": 215,
"line_end": 215,
"snippet": "**Important:** Each revision should be targeted. Don't rewrite everything β fix what the reviewer fl"
},
{
"id": "blocker:SKILL.md:261:system-reconnaissance",
"category": "blocker",
"severity": "low",
"pattern": "System reconnaissance",
"file": "SKILL.md",
"line_start": 261,
"line_end": 261,
"snippet": "User already did work or you already did work. Just run the review loop on what exists."
},
{
"id": "blocker:SKILL.md:314:system-reconnaissance",
"category": "blocker",
"severity": "low",
"pattern": "System reconnaissance",
"file": "SKILL.md",
"line_start": 314,
"line_end": 314,
"snippet": "- The reviewer subagent has no memory of prior loops β include prior feedback context explicitly in "
},
{
"id": "blocker:SKILL.md:140:network-reconnaissance",
"category": "blocker",
"severity": "low",
"pattern": "Network reconnaissance",
"file": "SKILL.md",
"line_start": 140,
"line_end": 141,
"snippet": "- Receive only the **work product** (files, diffs) β not your thought process"
},
{
"id": "blocker:SKILL.md:159:network-reconnaissance",
"category": "blocker",
"severity": "low",
"pattern": "Network reconnaissance",
"file": "SKILL.md",
"line_start": 159,
"line_end": 160,
"snippet": "- Auth enforced; no unauthenticated access"
},
{
"id": "blocker:SKILL.md:316:network-reconnaissance",
"category": "blocker",
"severity": "low",
"pattern": "Network reconnaissance",
"file": "SKILL.md",
"line_start": 316,
"line_end": 317,
"snippet": "- This skill does not replace human code review for security-critical or compliance-sensitive code; "
}
],
"finding_verdicts": [
{
"id": "external_commands:SKILL.md:25:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 0.99,
"reason": "Line 25 uses Markdown backticks to format example user phrases. It contains no Ruby expression, shell command, or execution instruction."
},
{
"id": "external_commands:SKILL.md:94:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 0.99,
"reason": "Lines 94-99 are a fenced Markdown block containing example requests. The backticks delimit documentation and do not execute commands."
},
{
"id": "external_commands:SKILL.md:99:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 0.99,
"reason": "Line 99 closes the Markdown example block opened on line 94. No command or executable Ruby syntax is present."
},
{
"id": "external_commands:SKILL.md:114:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 0.99,
"reason": "Lines 114-126 define a Mermaid flowchart inside a Markdown fence. The diagram documents control flow and does not invoke a shell."
},
{
"id": "external_commands:SKILL.md:126:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 0.99,
"reason": "Line 126 is the closing fence for the Mermaid diagram. It is Markdown syntax without execution behavior."
},
{
"id": "external_commands:SKILL.md:146:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 0.99,
"reason": "Lines 146-188 contain a reviewer prompt template in a fenced block. The fence is documentation syntax, not Ruby or shell execution."
},
{
"id": "external_commands:SKILL.md:188:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 0.99,
"reason": "Line 188 closes the documented reviewer prompt template. It does not contain an external command."
},
{
"id": "external_commands:SKILL.md:198:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 0.99,
"reason": "Lines 198-201 show a plain-text status format inside a Markdown fence. No executable content or command invocation appears."
},
{
"id": "external_commands:SKILL.md:201:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 0.99,
"reason": "Line 201 closes the status-format example. The backticks are Markdown delimiters only."
},
{
"id": "external_commands:SKILL.md:248:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 0.99,
"reason": "Lines 248-257 provide a prose walkthrough of a review cycle in a fenced block. They do not direct shell execution."
},
{
"id": "external_commands:SKILL.md:257:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 0.99,
"reason": "Line 257 closes a Markdown example block. It has no executable semantics."
},
{
"id": "external_commands:SKILL.md:263:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 0.99,
"reason": "Lines 263-272 contain a prose example about reviewing an authentication module. The Markdown fence does not execute its text."
},
{
"id": "external_commands:SKILL.md:272:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 0.99,
"reason": "Line 272 closes the review example block. There is no Ruby code, shell syntax, or external command."
},
{
"id": "external_commands:SKILL.md:295:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 0.99,
"reason": "Lines 295-304 begin a sample review report inside a Markdown fence. The content is presentation text, not executable code."
},
{
"id": "external_commands:SKILL.md:304:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 0.99,
"reason": "Line 304 uses inline Markdown backticks around a variable name within a sample report. It does not execute the variable or a command."
},
{
"id": "blocker:SKILL.md:114:system-reconnaissance",
"verdict": "false_positive",
"confidence": 0.99,
"reason": "Lines 114-115 start a Mermaid diagram of the review workflow. They contain no system discovery or reconnaissance instruction."
},
{
"id": "blocker:SKILL.md:215:system-reconnaissance",
"verdict": "false_positive",
"confidence": 0.99,
"reason": "Line 215 advises targeted revisions based on reviewer feedback. It does not request system information or discovery activity."
},
{
"id": "blocker:SKILL.md:261:system-reconnaissance",
"verdict": "false_positive",
"confidence": 0.99,
"reason": "Line 261 describes reviewing an existing work product. It does not instruct the agent to inspect the host system."
},
{
"id": "blocker:SKILL.md:314:system-reconnaissance",
"verdict": "false_positive",
"confidence": 0.99,
"reason": "Line 314 explains that reviewer subagents lack prior-loop memory. It contains no reconnaissance request or system command."
},
{
"id": "blocker:SKILL.md:140:network-reconnaissance",
"verdict": "false_positive",
"confidence": 0.99,
"reason": "Lines 140-141 limit reviewer context to the work product and request actionable feedback. They do not probe networks or external services."
},
{
"id": "blocker:SKILL.md:159:network-reconnaissance",
"verdict": "false_positive",
"confidence": 0.99,
"reason": "Lines 159-160 list authentication and query checks as review criteria. They describe defensive code review, not network reconnaissance."
},
{
"id": "blocker:SKILL.md:316:network-reconnaissance",
"verdict": "false_positive",
"confidence": 0.99,
"reason": "Lines 316-317 state that human review remains necessary for sensitive code. This is a limitation, not a network discovery instruction."
}
],
"semantic_findings": [],
"critical_findings": [],
"high_findings": [],
"medium_findings": [],
"low_findings": [],
"dangerous_patterns": [],
"files_scanned": 1,
"total_lines": 333,
"audit_model": "codex",
"audited_at": "2026-08-06T01:47:57.641Z",
"agent_auto_install_policy": "allowed",
"manual_install_policy": "allowed"
},
"content": {
"user_title": "Improve Work Through Structured Review Loops",
"value_statement": "First drafts often miss requirements, edge cases, or clarity problems. This skill coordinates scored reviews and targeted revisions until a quality gate is met.",
"seo_keywords": [
"Claude",
"Codex",
"Claude Code",
"AI review loop",
"iterative code review",
"subagent reviewer",
"quality gate",
"document review",
"specification review"
],
"actual_capabilities": [
"Runs a worker and reviewer cycle for code, documents, specifications, and other work products.",
"Requests a numeric score and specific feedback from a separate reviewer context.",
"Applies critical and important feedback through targeted revisions.",
"Supports custom minimum loops, maximum loops, quality gates, and reviewer capabilities.",
"Reports each loop score, key feedback, final score, and completed loop count."
],
"limitations": [
"Requires subagent support or a separate fresh conversation for independent review.",
"Reviewer scores can be inconsistent when acceptance criteria are vague.",
"Multiple review loops increase model usage, cost, and completion time.",
"Human review remains necessary for security-critical or compliance-sensitive work."
],
"use_cases": [
{
"title": "Harden a Feature Before Merge",
"description": "Review an implementation for correctness, edge cases, security issues, and maintainability before final delivery.",
"target_user": "Software engineers"
},
{
"title": "Refine a Technical Specification",
"description": "Identify unclear requirements, missing tradeoffs, and implementation gaps through repeated structured critique.",
"target_user": "Product managers and technical leads"
},
{
"title": "Polish Professional Documentation",
"description": "Improve structure, clarity, completeness, and audience fit while preserving the document's intended meaning.",
"target_user": "Technical writers"
}
],
"prompt_templates": [
{
"title": "Review Existing Work",
"prompt": "Run review-loop on this work. Use the default quality gate and show the score and key changes after each loop.",
"scenario": "Beginner review of an existing work product"
},
{
"title": "Implement and Review",
"prompt": "Implement the requested feature with review-loop. Require at least two loops, use a quality gate of eight, and summarize the final result.",
"scenario": "Standard feature implementation"
},
{
"title": "Polish a Design Document",
"prompt": "Review this design document for completeness, feasibility, alternatives, risks, and measurable outcomes. Run up to four loops with a quality gate of nine.",
"scenario": "Detailed technical document review"
},
{
"title": "Audit a Complex Change",
"prompt": "Use an advanced reviewer for this multi-file change. Check correctness, compatibility, security, failure modes, tests, and maintainability. Continue until gate nine or five loops.",
"scenario": "Advanced review with strict criteria"
}
],
"output_examples": [
{
"input": "Review an API endpoint with a quality gate of eight.",
"output": [
"Loop 1 of 4: Score 6/10. Missing input validation and timeout handling.",
"Loop 2 of 4: Score 8/10. Required fixes are complete; one naming issue remains.",
"Result: Quality gate met after two loops."
]
},
{
"input": "Polish a deployment specification using three review loops.",
"output": "The final specification scored 9/10 after three loops. Revisions clarified rollback steps, ownership, monitoring thresholds, and failure recovery."
}
],
"best_practices": [
"Define measurable acceptance criteria before starting the first review.",
"Give the reviewer only the relevant work product and task-specific criteria.",
"Track prior feedback so later revisions do not reintroduce resolved issues."
],
"anti_patterns": [
"Do not use repeated review loops for trivial edits with clear automated checks.",
"Do not accept a high score when the reviewer provides vague or unsupported feedback.",
"Do not continue beyond the maximum loop count without user approval."
],
"faq": [
{
"question": "What types of work can this skill review?",
"answer": "It can review code, specifications, design documents, refactors, bug fixes, infrastructure changes, tests, and professional writing."
},
{
"question": "Does this skill require subagent support?",
"answer": "Subagent support is preferred. A separate fresh conversation can provide an independent reviewer when subagents are unavailable."
},
{
"question": "What is the default quality gate?",
"answer": "The default gate is 8/10, with at least two loops and no more than four loops."
},
{
"question": "Can I change the loop settings?",
"answer": "Yes. Specify the minimum loops, maximum loops, quality gate, or reviewer capability in your request."
},
{
"question": "Will the loop always reach the target score?",
"answer": "No. Unclear requirements, weak criteria, or model limitations can prevent convergence before the maximum loop count."
},
{
"question": "Does this replace human code review?",
"answer": "No. Human review remains important for security, compliance, architecture, and other high-impact decisions."
}
]
},
"file_structure": [
{
"name": "SKILL.md",
"type": "file",
"path": "SKILL.md",
"lines": 333
}
]
}
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why β repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" β or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself β your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |