Apply a conservative fallback with a 15-tool-call budget for unmatched tasks.
npx skills add https://github.com/FRIKKern/doey --skill unknown-task
Please help me install this Agent Skill. Skill: unknown-task Source: https://github.com/FRIKKern/doey/tree/main/.claude/skills/unknown-task Command: npx skills add https://github.com/FRIKKern/doey --skill unknown-task
Provides a conservative fallback for tasks that do not match any specific skill, enabling safe execution when the exact skill is unknown.
When an input task has no matching skill, apply this fallback with a strict 15-tool-call budget and conservative edits.
High-intent search queries and answers about installing and using this skill.
A tool-call budget limits AI execution by capping the maximum number of tool invocations, up to 15 calls, during unknown tasks. This conservative execution strategy minimizes risk and prevents runaway actions when no matching skill is found.
A tool-call budget limits AI execution by capping the maximum number of tool invocations, up to 15 calls, during unknown tasks. This conservative execution strategy minimizes risk and prevents runaway actions when no matching skill is found.
Explicitly stating assumptions before acting provides clear context for ambiguous tasks, ensuring the AI operates safely under strict tool-call budgets. This conservative approach prevents destructive actions and promotes minimal, verifiable edits when handling unknown tasks.
Explicitly stating assumptions before acting provides clear context for ambiguous tasks, ensuring the AI operates safely under strict tool-call budgets. This conservative approach prevents destructive actions and promotes minimal, verifiable edits when handling unknown tasks.
This fallback approach limits AI to a maximum of 15 tool calls per task and requires stating assumptions before acting. It is designed for conservative execution of unknown tasks and does not support or permit destructive actions.
This fallback approach limits AI to a maximum of 15 tool calls per task and requires stating assumptions before acting. It is designed for conservative execution of unknown tasks and does not support or permit destructive actions.
reveniumEnforce Revenium token budgets before OpenClaw agent tool calls.
ToXMonEnforce conservative coding practices and surgical edits for LLM-generated changes.
juanreInvoke JSON Schema-defined tools across OpenAI, Anthropic, Gemini, and Ollama.
twkevinzhangReorganize task implementation plans into parallel execution structures with dependencies and agent assignments.
