The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external content such as package metadata, changelogs, and migration guides.
Ingestion points: The skill reads output from npm view, content from CHANGELOG.md or MIGRATION.md files on GitHub, and web search results for upgrade instructions.
Boundary markers: No specific delimiters or instructions to ignore embedded malicious directives in external content are defined.
Capability inventory: The agent has the ability to execute npm install (modifying the file system), npm run build/test (executing shell commands), and grep (reading local files).
Sanitization: No explicit sanitization or filtering of external text is performed before processing.
[COMMAND_EXECUTION]: The skill uses standard CLI tools to manage the development environment.
Commands: Executes npm ls, npm view, npm install, npm run build, npm test, npm run lint, npx tsc, and grep to analyze and update the project.
[EXTERNAL_DOWNLOADS]: Fetches package data and documentation from well-known services.
Sources: Downloads package information and payloads from the official npm registry and retrieves release notes from GitHub repositories and public web searches.

updating-npm-package