The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the user to install the cli-anything-firefly-iii package via pip from PyPI. This is a vendor-owned resource associated with the author and serves as the primary executable component of the skill.
[COMMAND_EXECUTION]: The skill operates by executing shell commands (e.g., cli-anything-firefly-iii accounts list) to interact with the Firefly III API. It provides a wide range of CRUD (Create, Read, Update, Delete) capabilities over financial data.
[INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it processes untrusted data from external financial records.
Ingestion points: Data is ingested through commands like transactions list, accounts get, and search transactions which retrieve text fields such as descriptions, notes, and category names from the Firefly III instance.
Boundary markers: No explicit delimiters or boundary markers (like XML tags or specific guardrail prompts) are defined in the instructions to separate data from instructions.
Capability inventory: The agent can execute any of the defined CLI commands, potentially allowing a malicious transaction description to influence future financial operations (e.g., "Delete account 123" embedded in a note).
Sanitization: No sanitization or validation logic for ingested string data is specified in the skill instructions.
[DATA_EXPOSURE]: The skill handles sensitive Personal Access Tokens (PAT) and financial data. It follows security best practices by recommending the use of environment variables for credentials rather than passing them as plain-text command-line arguments.

cli-anything-firefly-iii