The Agent Skills Directory

[DYNAMIC_CONTEXT_INJECTION]: The skill contains a shell command executed via the !command`` syntax at load time in SKILL.md. This command uses Node.js to read and parse the local package.json file, extracting dependency versions and script names to provide the agent with a snapshot of the current project state.
[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it processes untrusted data from the local project environment.
Ingestion points: Reads project metadata from package.json (via load-time script) and searches the codebase (using rg) for patterns like LiveUpdate and capacitor-live-update in files like src, ios, android, and configuration files.
Boundary markers: Absent. The data is interpolated directly into the context without delimiters or warnings to ignore embedded instructions.
Capability inventory: The skill has the capability to perform package management (npm install/uninstall), synchronize Capacitor projects (npx cap sync), and interact with external services via the Capgo CLI (npx @capgo/cli).
Sanitization: Absent. Content from the file system is not sanitized before being presented to the agent.
[COMMAND_EXECUTION]: The skill provides instructions for the agent to execute multiple shell commands, including npm, npx, and rg, to perform the migration and verification steps. These commands are consistent with the skill's stated purpose.

capawesome-live-update-migration