70High
The skill is vulnerable to prompt injection, allowing sub-
Behavior Analysis
Description verified — behavior matches claim
Claims to do
Feature Development: You are helping a developer implement a new feature. Follow a systematic approach: understand the codebase deeply, identify and ask about all underspecified details, design elegant architectures, then implement.
Actually does
The skill orchestrates a multi-phase feature development process by launching specialized AI agents (feature-description, code-explorer, code-architect, code-reviewer) to gather information, understand code, design, and review. It accesses and updates `IMPLEMENTATION.md` and `CLAUDE.md` files, reads from user-provided tickets/libraries (potentially via `MCP`), and uses `TodoWrite` for progress tracking.
Skill Info
NameFleron/Claude-plugins/extensive-build
Registrygithub
Version5662b3e1543f
PURLpkg:github/Fleron/Claude-plugins@5662b3e1543f?skill=extensive-build
Installs455,628
Source
SHA-2566bfddea93d25...
SHA-16ad8d023a128...
MD577ec8350aa7a...
TLSH52e1f926e7c0...
ssdeep192:t2Y/Znba...
Size6,827 bytes
Install
This skill has been flagged as potentially malicious. Review the findings below before installing.
Assessments (1)
AI Agent Traps ↗Prompt injection via sub-agent leading to file readhighreconnaissance
The skill launches sub-agents with templated prompts that can be influenced by user input. A malicious user could inject directives into these templates to instruct sub-agents to identify sensitive files, which the main agent is then explicitly instructed to read.
90% confidenceline 70
Example agent prompts: - 'Find features similar to [feature]...'; 'Once the agents return, please read all files identified by agents to build deep understanding'
File/system write for persistence or data exfiltrationmediumpersistence
The skill is instructed to update `IMPLEMENTATION.md`, `CLAUDE.md`, and external ticket systems. If an attacker can inject malicious content into the summaries or outputs that are written, this could be used for persistence (e.g., embedding malicious instructions in documentation) or data exfiltration.
80% confidenceline 30
'Extend it if new information is coming up.' (IMPLEMENTATION.md); 'ENSURE THE TICKET HAS BEEN UPDATED...WITH THE OUTPUT'; 'Update any CLAUDE.md and IMPLEMENTATION.md files accordingly'
Badge
Markdown
[](https://mondoo.com/ai-agent-security/skills/github/Fleron/Claude-plugins/extensive-build)HTML
<a href="https://mondoo.com/ai-agent-security/skills/github/Fleron/Claude-plugins/extensive-build"><img src="https://mondoo.com/ai-agent-security/api/badge/github/Fleron/Claude-plugins/extensive-build.svg" alt="Mondoo Skill Check" /></a>Image URL
https://mondoo.com/ai-agent-security/api/badge/github/Fleron/Claude-plugins/extensive-build.svgSecure 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