rdleclerc Bad Mini Skill
This fixture is intentionally bad. It has a vague trigger, a pinned model, no completion discipline, and a machine-specific local path: /Users/example/private/workspace.
The preamble is bloated by design. It repeats ordinary process filler that does not change what an agent should do. It adds ceremony without a sharp boundary. It asks the agent to be generally helpful. It includes setup chatter, migration chatter, review chatter, collaboration chatter, and background chatter without giving a concrete task class. It tells the reader to think carefully, be thoughtful, use judgment, consider tradeoffs, respect the user, keep context in mind, and make good choices, but it never names the observable situation that should activate the skill. It does not separate trigger, do, fallback, and invalid cases. It does not route resources. It does not state a side-effect boundary. It does not tell the reader what done means.
The preamble is bloated by design. It repeats ordinary process filler that does not change what an agent should do. It adds ceremony without a sharp boundary. It asks the agent to be generally helpful. It includes setup chatter, migration chatter, review chatter, collaboration chatter, and background chatter without giving a concrete task class. It tells the reader to think carefully, be thoughtful, use judgment, consider tradeoffs, respect the user, keep context in mind, and make good choices, but it never names the observable situation that should activate the skill. It does not separate trigger, do, fallback, and invalid cases. It does not route resources. It does not state a side-effect boundary. It does not tell the reader what done means.
The preamble is bloated by design. It repeats ordinary process filler that does not change what an agent should do. It adds ceremony without a sharp boundary. It asks the agent to be generally helpful. It includes setup chatter, migration chatter, review chatter, collaboration chatter, and background chatter without giving a concrete task class. It tells the reader to think carefully, be thoughtful, use judgment, consider tradeoffs, respect the user, keep context in mind, and make good choices, but it never names the observable situation that should activate the skill. It does not separate trigger, do, fallback, and invalid cases. It does not route resources. It does not state a side-effect boundary. It does not tell the reader what done means.
The preamble is bloated by design. It repeats ordinary process filler that does not change what an agent should do. It adds ceremony without a sharp boundary. It asks the agent to be generally helpful. It includes setup chatter, migration chatter, review chatter, collaboration chatter, and background chatter without giving a concrete task class. It tells the reader to think carefully, be thoughtful, use judgment, consider tradeoffs, respect the user, keep context in mind, and make good choices, but it never names the observable situation that should activate the skill. It does not separate trigger, do, fallback, and invalid cases. It does not route resources. It does not state a side-effect boundary. It does not tell the reader what done means.
The preamble is bloated by design. It repeats ordinary process filler that does not change what an agent should do. It adds ceremony without a sharp boundary. It asks the agent to be generally helpful. It includes setup chatter, migration chatter, review chatter, collaboration chatter, and background chatter without giving a concrete task class. It tells the reader to think carefully, be thoughtful, use judgment, consider tradeoffs, respect the user, keep context in mind, and make good choices, but it never names the observable situation that should activate the skill. It does not separate trigger, do, fallback, and invalid cases. It does not route resources. It does not state a side-effect boundary. It does not tell the reader what done means.
Do things for the user.