The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes shell commands in SKILL.md to perform environment checks (node -v, tsc --version) and project validation using npm run, npx tsc, and npx vitest. These commands execute code defined within the local project environment.
[DYNAMIC_EXECUTION]: The included diagnostic script scripts/ts_diagnostic.py uses subprocess.run() with shell=True to perform system queries and run analysis tools like grep and tsc on the local codebase.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from local project files including package.json and tsconfig.json. This data is used to determine tooling and project structure, creating a surface where a malicious project could influence agent behavior.
Ingestion points: Reads package.json, tsconfig.json, and searches source files via grep in scripts/ts_diagnostic.py and SKILL.md.
Boundary markers: No explicit delimiters or warnings are used when the agent processes or displays data parsed from project configuration files.
Capability inventory: The skill possesses file system read access and shell command execution capabilities via npm, npx, and Python's subprocess module.
Sanitization: The skill does not implement specific sanitization for data extracted from package.json or other configuration files before using it in decision-making logic.
[EXTERNAL_DOWNLOADS]: Instructions in sub-skills/migration-expertise.md and sub-skills/cli-debugging-tools.md suggest the use of npx to run tools like typesync, ts-migrate, and tsx, which involves downloading packages from the public npm registry.

typescript-expert