The Agent Skills Directory

[COMMAND_EXECUTION]: The diagnostic script (scripts/ts_diagnostic.py) uses subprocess.run with shell=True to execute system commands. This method is utilized to enable shell-specific features like piping and output redirection, but it represents a security risk as it could allow for command injection if the strings being executed were manipulated by untrusted data.\n- [COMMAND_EXECUTION]: The instructions in SKILL.md recommend running destructive commands, specifically rm -rf node_modules/.cache .tsbuildinfo, to troubleshoot module resolution errors. While these are common development fixes, they involve significant file system modifications.\n- [EXTERNAL_DOWNLOADS]: The skill frequently uses bunx and npx to download and run various development tools (including tsc, ts-migrate, typesync, and vitest) from the public NPM registry during its diagnostic and maintenance tasks.\n- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it reads and processes local configuration files (package.json, tsconfig.json) and source code files from the project directory.\n
Ingestion points: Files are read using Python's open() function and scanned with system utilities like grep and tsc.\n
Boundary markers: There are no explicit delimiters or specific instructions to the agent to ignore potentially malicious directions found within the files it analyzes.\n
Capability inventory: The agent is authorized to run shell commands, perform file system operations, and delete project directories.\n
Sanitization: The skill does not perform validation or sanitization on the content of the files it reads before processing them as part of its diagnostic logic.

typescript-expert