Back to skill

Security audit

typescript-dev

Security checks for vulnerabilities and agentic risk

Overview

This TypeScript development guide fits its stated purpose, though its installer and configuration migration examples require care.

Prefer version-locked project tools over the download-to-Bash installer. Back up Tailwind configuration and test the migration before deleting it. Run deployment and optional skill/MCP setup only when you intend those changes. All bundled files were inspected; remotely downloaded packages and installer contents were outside this review.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Findings (13)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
92% confidence
Finding

The Vite build matches the stated build functionality. However, pnpm wrangler deploy performs an external deployment that can change hosted resources, rather than merely building or checking the project. That operational capability is not declared.

Content

No source excerpt is available for this finding.

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 24)May include surrounding context.

md
- **[references/shadcn.md](references/shadcn.md)** - shadcn/ui CLI, component authoring with CVA + `data-slot`, the `cn` package, registries, Base UI / Radix /

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 26)May include surrounding context.

md
- **[references/vitest.md](references/vitest.md)** - Vitest 5 config, Testing Library, mocking, coverage, browser mode, projects, v4->v5 migration, test speed.

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 27)May include surrounding context.

md
- **[references/hono.md](references/hono.md)** - Hono 4 web framework: routing, context, middleware, validation (Zod), end-to-end type-safe RPC, OpenAPI, helper

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 140)May include surrounding context.

md
[hono.md](references/hono.md).

External Script Fetching

High
Category
Supply Chain
Confidence
97% confidence
Finding

curl -fsSL https://biomejs.dev/install | bash immediately executes whatever script the endpoint returns, including after redirects. HTTPS protects transport but does not establish an immutable, reviewed artifact, so upstream compromise could become arbitrary local code execution; there is no evidence that the author intends such a compromise.

Content

Scanner excerpt · references/biome.md (reported line 271)May include surrounding context.

biome migrate stylelint --write # 2.6: import a Stylelint config

text

Standalone install without npm: `curl -fsSL https://biomejs.dev/install | bash` (2.6). 2.5.15 also restricts the daemon's Unix socket to the current user (`0600`, in a private `biome-daemon` cache dir) - another reason to stay on the latest patch. Biome 2.5 also adds `formatter.delimiterSpacing` (pads `[ 1, 2 ]` / `{ a }` / `( x )` - an accessibility aid for dyslexia; behavior varies per language) and can now format/lint `.svg` files.

## Gotchas

The skill instructs deletion of an existing Tailwind configuration without warning about loss of project-specific settings.

Medium
Category
Output Handling
Confidence
88% confidence
Finding

The instruction to delete an existing tailwind.config.js directly affects user-owned project configuration. Although it also calls for migrating values into CSS, it does not warn about losing settings or require preserving and validating the configuration before deletion.

Content

No source excerpt is available for this finding.

MCP server referenced without pinned version: 'npx tsc'.

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.

The standalone installation instruction executes a downloaded script without warning the user about remote code execution.

Medium
Category
Output Handling
Confidence
95% confidence
Finding

The standalone installation example pipes a mutable remote script directly into Bash without inspection or integrity verification. If the delivery endpoint or upstream infrastructure is compromised, following this development instruction could execute arbitrary code with the agent's permissions and expose repository contents or credentials; this is the same underlying issue as the SC2 finding.

Content

No source excerpt is available for this finding.

MCP server referenced without pinned version: 'npx skills'.

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.

MCP server referenced without pinned version: 'npx @tailwindcss/upgrade'.

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.

Troubleshooting instructs deletion of the existing Tailwind configuration without warning about configuration loss.

Medium
Category
Output Handling
Confidence
96% confidence
Finding

The troubleshooting entry tells the agent to delete the configuration before running the migration tool, potentially removing the input needed to preserve custom themes, plugins, and source paths. Although earlier sections recommend migration, this actionable ordering can cause configuration loss and broken builds or styles when followed literally.

Content

No source excerpt is available for this finding.

MCP server referenced without pinned version: 'npx vitest'.

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.