Security audit
Security checks for vulnerabilities and agentic risk
This skill is a disclosed design-to-code workflow that modifies frontend code and writes a plan report as part of its stated purpose, with no evidence of hidden execution, credential access, persistence, or exfiltration.
Install this if you want an agent to implement frontend UI from design artifacts. Expect it to inspect project components, use available design-tool integrations or screenshots, modify frontend files, run available checks, and create a timestamped design plan report in the repository.
Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.
- ID: `fec-implement-from-design`
- Category: `design-ui`
- Version: `2.9.0`
- Source: `skills/fec-implement-from-design/SKILL.md`
## Description
The README uses very broad activation language such as implementing UI from multiple design sources and generic design-to-code tasks, which can cause the skill to trigger on a wide range of normal frontend requests without clear scoping boundaries. In an agent ecosystem, over-broad routing can lead to unintended invocation, incorrect tool selection, and increased exposure to downstream prompt or workflow behaviors the user did not explicitly request.
The Chinese trigger examples are vague and semantically broad, so common phrases about "design draft" or "implementation by design" may activate the skill for ordinary requests. That increases the chance of misrouting user tasks into a workflow that assumes design-source ingestion and repository modification, which can produce incorrect or excessive changes.
The Chinese trigger examples are vague and semantically broad, so common phrases about "design draft" or "implementation by design" may activate the skill for ordinary requests. That increases the chance of misrouting user tasks into a workflow that assumes design-source ingestion and repository modification, which can produce incorrect or excessive changes.
The description activates on very broad phrases like implementing UI from many design sources, which can cause the skill to be invoked outside its intended boundaries. In agent systems, overbroad routing increases the chance that unrelated tasks, sensitive visual inputs, or unsafe design-to-code requests are automatically handled by this skill without explicit user intent or scope checks.
The package description is extremely broad and covers many ordinary frontend tasks such as implementing UI from design files, screenshots, and general design-to-code work. In agent-routing systems, this can cause over-triggering on common requests, leading the skill to activate outside its intended scope and potentially influence workflows where a more specific or safer skill should be used.
Line L14 explicitly calls out 'Chinese triggers include design draft, implementation by design,' which introduces a language-specific activation policy in the skill description. Because no opt-in, multilingual choice, or justification is provided here, this may conflict with language/locale neutrality expectations.
The skill directs the agent to save a timestamped report file into the repository but does not clearly disclose that this will create a new file. Unexpected writes are risky because they can modify the working tree, add noise to commits, or create artifacts the user did not explicitly request.
The description includes Chinese trigger phrases without stating that Chinese-language activation is optional or based on the user's locale and intent. This can cause unintended routing for multilingual users, where natural-language mentions of those phrases invoke the skill unexpectedly even when the user did not choose this workflow.
The manifest explicitly states 'Chinese triggers include design draft, implementation by design,' which introduces a language-specific invocation policy. There is no accompanying indication that language selection is optional or user-driven.
No suspicious patterns detected.