The Agent Skills Directory

[DYNAMIC_CONTEXT_INJECTION]: The skill uses the !command syntax in SKILL.md to execute a Node.js script at load time. This script identifies current @capacitor/ dependencies by reading the local package.json file. This provides the AI agent with immediate context for the upgrade. The command is benign as it performs a read-only operation on standard project metadata and does not incorporate external arguments or exfiltrate data.
[INDIRECT_PROMPT_INJECTION]: The skill processes data from package.json, creating a potential vulnerability surface where malicious content in that file could attempt to influence the agent's behavior.
Ingestion points: The package.json file is read and its dependency list is injected into the agent's prompt during the 'Live Project Snapshot' phase.
Boundary markers: Absent; the dependency list is presented to the agent without specific delimiters or instructions to treat the content as potentially untrusted data.
Capability inventory: The skill has access to powerful tools including npm, npx, and arbitrary node commands, which increases the potential impact of a successful injection.
Sanitization: The script includes a logical filter that only processes keys starting with @capacitor/, which significantly reduces the attack surface compared to reading the entire file, though it does not sanitize the version strings themselves.

capacitor-app-upgrade-v5-to-v6