The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to source the user's shell profiles (~/.zshenv and ~/.bashrc) at the start of a session to ensure the PATH is correctly configured for the clams CLI. This is a common practice for CLI integrations to ensure the binary is discoverable.
[COMMAND_EXECUTION]: The skill includes several shell scripts in the scripts/ directory for rendering PDF reports and verifying the CLI state. These scripts execute standard command-line utilities such as jq, awk, and weasyprint to process JSON data from the CLI.
[INDIRECT_PROMPT_INJECTION]: The skill processes external data via CSV imports (clams connections import) and custom mapping configurations.
Ingestion points: CSV files provided by the user and processed by the clams CLI.
Boundary markers: Not explicitly defined in the prompts, but the skill relies on structured JSON output (--machine --format json) for script processing.
Capability inventory: The skill can execute CLI commands, write to specified output paths, and run internal shell scripts.
Sanitization: Data parsing and validation are handled by the underlying clams binary.
[DATA_EXPOSURE]: The skill handles sensitive financial information (balances, transactions, cost basis). The README.md contains a dedicated privacy section informing users that this data is sent to the AI model provider during use.

clams