The Agent Skills Directory

[DYNAMIC_CONTEXT_INJECTION]: The skill employs the dynamic context syntax to execute shell commands when the skill is loaded. Specifically, it runs a node script to parse the project's package.json and a find command to locate Capacitor configuration files. These commands are used to provide a snapshot of the current application state, enabling the agent to plan the migration steps accurately. The commands are limited to local file system inspection and do not involve network activity or access to sensitive credentials.
[INDIRECT_PROMPT_INJECTION]: The skill has an indirect prompt injection surface because it ingests data from the local project environment into the agent's context.
Ingestion points: The package.json file is read and processed by a script in SKILL.md via dynamic context injection.
Boundary markers: There are no explicit delimiters or instructions to the agent to ignore potentially malicious content within the processed file.
Capability inventory: The skill instructs the agent to use tools for package management (npm), Capacitor migration (npx cap migrate), and native builds (gradlew, xcodebuild) as described in the Procedure and Verification sections.
Sanitization: The ingestion script provides a layer of filtering by only extracting package names that begin with @capacitor/, which limits the scope of data processed from the package manifest.

capacitor-app-upgrades