--- name: llm-wakeup description: >- Use when the user asks to schedule or manage an automatic assistant-session wakeup after a usage reset or timeout. --- > Generated from `blueprint.yaml`. Do not edit this block by hand. Executable Interfaces: Send the required `caller` (caller skill), `interface`, `version`, and `arguments`; optional `dry_run` defaults to false. Compact uses ordered `positionals` plus an option mapping; ordered raw argv uses `positionals: []` plus every argv token in list `options`. Never mix forms. - `llm-wakeup._rtx.interface.auto-policy` — Manage opt-in automatic near-limit wakeup scheduling for one provider session. - Caller: `llm-wakeup` - Version: 1 - Security level: 2 - Alternative: `default` Arguments JSON (replace labels with actual values). Omit optional positionals and options that are not needed. {"options": {}, "positionals": ["action", "provider", "session-id"], "stdin": null} Required options: []; positional arity: 1..3; stdin: forbidden - `llm-wakeup._rtx.interface.explicit-schedule` — Persist a guarded wakeup for an explicitly identified provider session and reset time. - Caller: `llm-wakeup` - Version: 1 - Security level: 2 - Alternative: `default` Arguments JSON (replace labels with actual values). Omit optional positionals and options that are not needed. {"options": {"--delay": "duration", "--message": "message"}, "positionals": ["provider", "session-id", "reset-time"], "stdin": null} Required options: []; positional arity: 3..3; stdin: forbidden - `llm-wakeup._rtx.interface.infer-schedule` — Infer provider, canonical session, and reset time before persisting a guarded wakeup. - Caller: `llm-wakeup` - Version: 1 - Security level: 2 - Alternative: `default` Arguments JSON (replace labels with actual values). Omit optional positionals and options that are not needed. {"options": {"--delay": "duration", "--message": "message", "--text": "timeout-or-resume-text"}, "positionals": [], "stdin": null} Required options: []; positional arity: 0..0; stdin: forbidden - `llm-wakeup._rtx.interface.setup` — Install or refresh the wakeup integration without caller arguments. - Caller: `llm-wakeup` - Version: 1 - Security level: 2 - Alternative: `default` Arguments JSON (replace labels with actual values). Omit optional positionals and options that are not needed. {"options": {}, "positionals": [], "stdin": null} Required options: []; positional arity: 0..0; stdin: forbidden When this skill is used, begin with: Skill: llm-wakeup ## Route by intent - Enable, disable, or inspect automatic wakeups: invoke `auto-policy`. The level is the action argument. `on` wakes the session only when the provider refused a turn for lack of quota and the session stopped there. `force` wakes it at reset whenever usage neared the limit, refused or not; pass it only when the user asks to be woken either way. `off` removes the policy, and `status` reports the level in effect. - Schedule after a timeout when provider, session, or reset may need discovery: invoke `infer-schedule`. - Schedule when provider, session, and reset are all explicit: invoke `explicit-schedule`. - Set up, refresh, or tear down the optional wakeup commands and dedicated due-delivery registration: invoke `setup`. For setup, first obtain the exact selected Python fingerprint and current plugin root; do not discover Python. Pass `message` and `delay` only when the user supplies them. Otherwise preserve the interface defaults. Do not guess a provider, session, alias, or reset time after an interface reports ambiguity. For `auto-policy`, ask for the missing provider or canonical session and retry `auto-policy`. For `infer-schedule`, ask for the missing explicit value; when provider, session, and reset are all explicit, route to `explicit-schedule`. For `auto-policy`, report the resolved provider, canonical session, and policy state, including which level is in effect. For `infer-schedule` and `explicit-schedule`, report the resolved provider, canonical session, and local scheduled time. Report failures plainly without claiming a wakeup exists.