What problem does it solve? Agents routinely over-engineer content insertion by jumping to scripts when a simple form input or paste would work, and they mark tasks complete based on unreliable visual confirmation. This Skill replaces ad hoc judgment with a deterministic, testable decision procedure for inserting content into documents, web forms, and editors. ## Core Features & Use Cases - Ranked method selection: The decide() function always starts at the lowest-complexity eligible method (form input, typing, paste, upload) and only appends scripting as a last-resort fallback when the automation gate is open. - Pre-execution linting: lint_strict() enforces rules LINT-001 through LINT-008, blocking premature scripting, disabled verification, complexity bias, and unsafe content such as absolute workstation paths or mojibake. - Verified execution with fallback: execute_with_fallback() runs the chosen chain and confirms success only via a programmatic signature check, never visual confirmation, with a blocked-decision path that notifies the user when no method is eligible. - Use Case: An agent must insert a 2,000-character article into a CMS field. It calls decide() with the task and environment, receives direct_form_input with paste as fallback, lints the decision, executes, and verifies the signature in extracted page text before marking the task complete. ## Quick Start Ask the agent to decide and execute the correct insertion method for your content using the CIDF decide, lint, and execute_with_fallback API with a verification signature.
