The Agent Skills Directory

[DATA_EXFILTRATION]: The splitwise-pp-cli supports a --deliver webhook:<url> flag that can POST sensitive financial results, such as balances, spend history, and debts, to any external URL provided as an argument.
[REMOTE_CODE_EXECUTION]: The skill provides instructions for the agent to download and execute code at runtime using npx -y @mvanhorn/printing-press-library and go install github.com/mvanhorn/printing-press-library/....
[EXTERNAL_DOWNLOADS]: Core functionality depends on downloading and installing the splitwise-pp-cli and splitwise-pp-mcp binaries from external registries including npm and GitHub.
[COMMAND_EXECUTION]: The skill relies on executing shell commands via the Bash tool to interact with the Splitwise API and the local SQLite ledger using the splitwise-pp-cli binary.
[PROMPT_INJECTION]: The skill processes untrusted data from Splitwise (expense descriptions, comments, and group names) which can contain malicious instructions.
Ingestion points: Reads from the Splitwise API and local synced store via search, ledger, and activity commands.
Boundary markers: None identified in the prompt templates to distinguish between data and instructions.
Capability inventory: The agent has access to Bash for command execution and file system access.
Sanitization: No explicit sanitization or escaping of external financial data is described before it is returned to the agent context.

pp-splitwise