Review CLAUDE.md from conversation history

Name: Claude-plugins/review-claudemd
Author: Fleron

View on GitHub↗

Share on X Share on LinkedIn Share on Bluesky

40Medium

The skill risks resource exhaustion from

ThreatsCI SM CS BC S HITL

Behavior Analysis

Description verified — behavior matches claim

Claims to do

Review CLAUDE.md from conversation history: Analyze recent conversations to improve both global (~/.claude/CLAUDE.md) and local (project) CLAUDE.md files.

Actually does

The skill uses `pwd`, `sed`, `ls`, `date`, `mkdir`, `basename`, `cat`, `jq`, and `grep` to locate conversation history in `~/.claude/projects/`, extract the 20 most recent conversations from `.jsonl` files, and convert them into plain text files in a temporary directory (`/tmp/claudemd-review-*`). It then describes launching internal 'Sonnet subagents' to analyze these conversations along with `~/.claude/CLAUDE.md` and `./CLAUDE.md` to identify potential improvements, and finally aggregates these findings for the user.

Threat Analysis

AI Agent Traps ↗Scanned 4/15/2026

Content InjectionPerceptionclean

Semantic ManipulationReasoningclean

Cognitive StateMemory & Learning1 finding

Behavioural ControlAction1 finding

SystemicMulti-Agent Dynamicsclean

Human-in-the-LoopHuman Overseerclean

Skill Info

NameFleron/Claude-plugins/review-claudemd

Registrygithub

Version5662b3e1543f

PURLpkg:github/Fleron/Claude-plugins@5662b3e1543f?skill=review-claudemd

Installs455,628

Source

Repository ↗SKILL.md ↗

SHA-25695db0ae0b08f...

SHA-159739a60580c...

MD5945ac34fffb6...

TLSHec5166aa9061...

Size2,718 bytes

Install

This skill has been flagged as potentially malicious. Review the findings below before installing.

GitHub

https://github.com/Fleron/Claude-plugins/tree/main/skills/review-claudemd

455,628 installs

Claude CodeDocs

/plugin marketplace add Fleron/Claude-plugins

/plugin install review-claudemd@Fleron/Claude-plugins

Skills.shDocs

npx skills add https://github.com/Fleron/Claude-plugins --skill review-claudemd

Assessments (2)

AI Agent Traps ↗

Sub-agent spawning and potential resource abusemediumsub agent spawning

The skill explicitly launches 'parallel Sonnet subagents' to process data. While the subagent prompt is fixed, the number of agents is not explicitly capped, which could lead to resource exhaustion or amplify other attacks if not properly managed by the underlying platform.

90% confidenceline 48

Launch parallel Sonnet subagents to analyze conversations.

RAG/Knowledge base poisoning via suggested editslowrag poisoning

The skill's primary function is to suggest improvements to `CLAUDE.md` files, which serve as a knowledge base for the agent. Malicious or subtly biased suggestions, if approved by a human, could poison the agent's future instructions or context, leading to long-term behavioral changes.

80% confidence

Analyze the conversations against BOTH CLAUDE.md files. Find: ... Patterns that should be added to LOCAL CLAUDE.md ... Patterns that should be added to GLOBAL CLAUDE.md ... Ask user if they want edits drafted.

Badge

Markdown

[![Mondoo Skill Check](https://mondoo.com/ai-agent-security/api/badge/github/Fleron/Claude-plugins/review-claudemd.svg)](https://mondoo.com/ai-agent-security/skills/github/Fleron/Claude-plugins/review-claudemd)

HTML

<a href="https://mondoo.com/ai-agent-security/skills/github/Fleron/Claude-plugins/review-claudemd"><img src="https://mondoo.com/ai-agent-security/api/badge/github/Fleron/Claude-plugins/review-claudemd.svg" alt="Mondoo Skill Check" /></a>

Image URL

https://mondoo.com/ai-agent-security/api/badge/github/Fleron/Claude-plugins/review-claudemd.svg

Secure your AI agents

Skills can read files, run commands, and access credentials. Mondoo helps organizations manage the security risks of AI agent skills across their entire fleet.

Continuous scanning of skills across all registries
Policy enforcement before skills reach your agents
Integration with your existing security workflow