The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from external files such as CSV, Excel, and JSON provided by the user, which presents a surface for indirect prompt injection.
Ingestion points: instructions.md specifies that the agent should accept data in common formats (CSV, Excel, JSON) or via direct user input for processing.
Boundary markers: The instructions explicitly direct the agent to "Only use data provided by the user or referenced in the skill" and "Validate inputs before processing."
Capability inventory: The skill utilizes network access (via Python requests) and filesystem access (as defined in claw.json) to perform its automation tasks.
Sanitization: The agent is instructed to report errors clearly and validate inputs, though specific sanitization logic for the data content is not detailed in the script code.
[COMMAND_EXECUTION]: The skill includes Python scripts that utilize the requests library to interact with n8n webhooks and external APIs.
Evidence: SKILL.md provides a trigger_daily_report function that sends project data to a user-specified webhook URL via HTTP POST.
[SAFE]: All external domains referenced in the skill, including datadrivenconstruction.io, n8n.io, and openweathermap.org, are either official vendor resources or well-known, trusted service providers. The use of environment variables for API keys and service URLs follows security best practices.

n8n-daily-report