The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes several powerful CLI tools to perform its tasks.
It utilizes a specialized CLI tool called denvig (provided by the author) to identify outdated dependencies.
It performs standard project management tasks using pnpm install, git, and the GitHub CLI (gh).
These commands are restricted to the current project context and are necessary for the skill's primary function of dependency management.
[EXTERNAL_DOWNLOADS]: The skill interacts with external sources to gather information and update the project.
It fetches package metadata from the npm registry using npm view.
It retrieves release notes and changelogs from GitHub repositories via the WebFetch tool.
It downloads and installs third-party code packages through pnpm install as part of the upgrade process.
[INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection as it processes untrusted data from external sources.
Ingestion points: Content is ingested from GitHub changelog files and release pages via WebFetch (SKILL.md).
Boundary markers: There are no explicit delimiters or instructions provided to the agent to disregard instructions that might be embedded within the changelog text.
Capability inventory: The agent has the ability to run pnpm install, modify files, commit changes, and push to remote repositories.
Sanitization: There is no explicit sanitization or validation of the fetched changelog content before it is processed by the AI.
While this is a theoretical risk where a malicious package author could attempt to influence the agent's behavior through changelog content, it is documented here as an inherent risk of the task rather than a malicious finding.

denvig-upgrade-npm-dependencies