The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes npm install, npm run verify, npm run build, and npm test as part of the upgrade and verification process. These commands run scripts defined by the project author in package.json, which could lead to arbitrary code execution if the project being upgraded is malicious.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes data from the local workspace's package.json file to determine its upgrade strategy.
Ingestion points: The skill uses a node -e script to read and parse the package.json file from the current working directory.
Boundary markers: No explicit delimiters or instructions to ignore embedded content are used when processing the file data.
Capability inventory: The skill has the capability to execute shell commands (npm, npx, find), read files, and perform network operations via package managers.
Sanitization: No sanitization or validation of the ingested package.json content is performed before it is used to influence the agent's actions.
[EXTERNAL_DOWNLOADS]: The skill triggers npm install, which downloads dependencies from the public npm registry based on the definitions in the project's manifest files.

capacitor-plugin-upgrades