Security audit
Security checks for vulnerabilities and agentic risk
This skill is a coherent editable diagram workflow with disclosed local file, optional network, and output-writing behavior.
Install this for editable technical diagrams, but run code-structure scanners only on the intended project directory. For private architecture or regulated content, prefer local .drawio files and avoid diagrams.net URLs or CDN-rendered brand icons unless that data boundary is acceptable.
Referenced artifact was not completely inspected
- When you need fast browser opening or sharing, use [diagram-url.mjs](scripts/diagram-url.mjs) to generate the URL; when you need local auditable delivery, alw
Referenced artifact was not completely inspected
- When you need fast browser opening or sharing, use [diagram-url.mjs](scripts/diagram-url.mjs) to generate the URL; when you need local auditable delivery, alw
Referenced artifact was not completely inspected
structure graphs, first generate graph JSON and then use [layout-graph.mjs](scripts/layout-graph.mjs) for automatic layout.
Referenced artifact was not completely inspected
- Code structure diagram can be used first [scan-js-modules.mjs](scripts/scan-js-modules.mjs), [scan-ts-modules.mjs](scripts/scan-ts-modules.mjs), [scan-python-
Referenced artifact was not completely inspected
st [scan-js-modules.mjs](scripts/scan-js-modules.mjs), [scan-ts-modules.mjs](scripts/scan-ts-modules.mjs), [scan-python-modules.mjs](scripts/scan-python-modules
Referenced artifact was not completely inspected
scan-ts-modules.mjs](scripts/scan-ts-modules.mjs), [scan-python-modules.mjs](scripts/scan-python-modules.mjs), [scan-go-packages.mjs](scripts/scan-go-packages.m
Referenced artifact was not completely inspected
scan-go-packages.mjs](scripts/scan-go-packages.mjs), [scan-rust-modules.mjs](scripts/scan-rust-modules.mjs) or [scan-python-classes.mjs](scripts/scan-python-cla
Referenced artifact was not completely inspected
ust-modules.mjs](scripts/scan-rust-modules.mjs) or [scan-python-classes.mjs](scripts/scan-python-classes.mjs) to generate graph JSON.
Referenced artifact was not completely inspected
Tool script sharing [studio-core.mjs](scripts/studio-core.mjs). Offline query uses [shape-index.json](data/shape-index.json) and [brand-icons.json](data/brand-i
Referenced artifact was not completely inspected
tudio-core.mjs](scripts/studio-core.mjs). Offline query uses [shape-index.json](data/shape-index.json) and [brand-icons.json](data/brand-icons.json). Style pres
Referenced artifact was not completely inspected
ine query uses [shape-index.json](data/shape-index.json) and [brand-icons.json](data/brand-icons.json). Style preset resources include [schema.json](styles/sche
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-drawio-studio`
- Category: `design-ui`
- Version: `2.9.0`
- Source: `skills/fec-drawio-studio/SKILL.md`
## Description
The skill references network-capable behavior such as generating diagrams.net viewer/editor URLs, using networked branding icons, and interacting with an external draw.io MCP ecosystem, but it does not declare any explicit tool scope or allowed-tools restrictions. In an agent environment, this can enable unintended outbound access or data transfer without clear authorization boundaries, which is risky because diagram content may contain sensitive architecture or system information.
The description and trigger language are broad enough to activate on many generic diagram-related requests without clear guardrails, which can cause the skill to be selected outside its intended boundary. In an agent setting, over-broad activation increases the chance of inappropriate tool use, unnecessary access to referenced scripts/resources, and misrouting tasks to a workflow with capabilities the user did not explicitly request.
This code defines d(e,r) to create directories and write files, and F(e,r) to write output to a provided path when r is set. There is no confirmation prompt, warning comment, docstring, or visible user-facing log indicating that the skill may modify the filesystem.
The guidance says to 'Use the same language as the user's request,' which is a natural-language locale policy embedded in the skill content. Under the stated policy, language constraints should either offer user choice or be clearly justified; this line imposes a language behavior without that framing.
No suspicious patterns detected.