The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill references several external resources and tools.
A documentation link to basarat.gitbook.io/typescript/ was flagged by automated scanners (URLite) as a phishing risk. While this is a well-known community resource, the domain reputation of GitBook subdomains can trigger alerts.
The instructions recommend using npx to run @typescript/analyze-trace and type-coverage. These commands download and execute code from the npm registry at runtime.
[COMMAND_EXECUTION]: The skill provides numerous shell commands for the agent to execute in a developer environment.
It suggests running tsc with various flags for diagnostics and performance tracing.
It includes a command to execute a local Node.js script: node --max-old-space-size=8192 ./node_modules/typescript/lib/tsc.js --noEmit. This relies on a relative path to the node_modules directory, which could be a vector for local code execution if the environment is not secured.
[INDIRECT_PROMPT_INJECTION]: The skill creates a surface for indirect prompt injection attacks.
Ingestion points: The agent is designed to analyze and process complex TypeScript code provided by users or read from project files (SKILL.md).
Boundary markers: The instructions do not define clear delimiters or "ignore" instructions for the code being analyzed, which could allow malicious instructions within the TypeScript comments or strings to influence agent behavior.
Capability inventory: The agent has the capability to execute shell commands (tsc, node, npx) based on its analysis of the code.
Sanitization: There are no mentioned mechanisms for sanitizing or validating the code before it is processed or used in diagnostic commands.

typescript-type-expert