Update to Date App Plugin

{
  "name": "dating-response-strategy",
  "description": "Convert grounded Match, Wiki, timeline, event, profile, and router briefs into an organic dating-message strategy using reciprocity, cognitive load, topic momentum, calibrated self-disclosure, humor, curiosity, and consent. Use after routing and before generating or reviewing message candidates.",
  "included_files": [
    {
      "relative_path": "agents/openai.yaml",
      "size_in_bytes": 214
    },
    {
      "relative_path": "references/response-principles.md",
      "size_in_bytes": 7680
    }
  ],
  "skill_md_contents": "---\nname: dating-response-strategy\ndescription: Convert grounded Match, Wiki, timeline, event, profile, and router briefs into an organic dating-message strategy using reciprocity, cognitive load, topic momentum, calibrated self-disclosure, humor, curiosity, and consent. Use after routing and before generating or reviewing message candidates.\n---\n\n# Dating Response Strategy\n\nChoose what the next message should accomplish and how it should be shaped. Do\nnot draft final copy in this skill.\n\nRead [references/response-principles.md](references/response-principles.md)\nbefore selecting a strategy or comparing candidate directions.\n\n## Required inputs\n\nConsume the relevant outputs of:\n\n- `interpret-match-context` for grounded facts and evidence classes;\n- `interpret-property-map` for private Wiki meanings and assessments;\n- `read-conversation-timeline` for the latest complete turn, active topics, and\n  response obligations;\n- `event-state-engine` for event boundaries and retry eligibility;\n- `personal-profile` for truthful self-reference and voice constraints;\n- `task-and-ruleset-router` for task, language, effort, ruleset, blockers, and\n  missing-context status.\n\nIf the router selected `pause_for_context`, `wait`, or `close`, honor it. Do not\ninvent a strategy that bypasses missing facts, event limits, or boundaries.\n\n## Strategic objective\n\nOptimize for an honest, organic exchange that makes the next conversational\nstep easy and helps both people reveal compatibility. Do not optimize for a\nreply at any cost, emotional dependency, control, extracting information, or\nperforming a tactic on the match.\n\nEvaluate a direction through six lenses:\n\n- `relevance`: it responds to what matters in the current turn;\n- `reciprocity`: its investment fits the interaction without mechanically\n  copying length or style;\n- `ease`: it is cognitively simple to understand and answer;\n- `distinctiveness`: it sounds specific enough not to be interchangeable with\n  any match;\n- `truthfulness`: every explicit and implied personal fact is supported;\n- `consent`: it respects boundaries, pressure limits, and event retry rules.\n\nTruthfulness and consent are hard gates. A direction that fails either is\ninvalid even if it seems engaging.\n\n## Choose the conversational move\n\nSelect one primary move:\n\n- `answer`: satisfy a direct question or practical decision;\n- `acknowledge`: show that a disclosure, emotion, joke, correction, or boundary\n  was received;\n- `relate`: add proportionate, verified self-disclosure;\n- `play`: continue a shared humorous or flirtatious premise;\n- `explore`: ask one specific question that grows from the active thread;\n- `clarify`: resolve material ambiguity without interrogation;\n- `callback`: revive a meaningful earlier detail or shared joke;\n- `bridge`: complete the current topic and enter a related or fresh one;\n- `discover`: surface one unresolved checklist or personality topic naturally;\n- `reenter`: open one low-pressure fresh thread when the event engine permits;\n- `advance`: offer one optional Instagram or Facebook exchange under CTA;\n- `close`: end respectfully;\n- `hold`: send nothing yet.\n\nAdd at most one secondary move when it makes the response more organic. A short\nreply may acknowledge then answer, answer then relate, or continue then bridge.\nDo not stack goals merely because the data contains them.\n\n## Shape the response\n\nUse this flexible architecture when helpful:\n\n1. `receive`: answer or acknowledge the most important current signal;\n2. `contribute`: add a specific reaction, true detail, interpretation, or\n   playful extension;\n3. `open`: leave an easy path forward through a question, unfinished thought,\n   choice, or natural statement.\n\nNot every message needs all three parts, and not every message needs a question.\nA relevant statement can invite response more naturally than repeated\ninterviewing.\n\nWhen several topics are active, satisfy every must-address obligation from the\ntimeline brief. Combine compatible topics, acknowledge important secondary\nones, and explicitly park only what cannot fit naturally. Never choose one topic\nat random and silently ignore questions or meaningful disclosures.\n\n## Apply messaging theory\n\n### Reciprocity\n\nCalibrate to her demonstrated engagement, not assumptions about interest. Match\nthe level of informational and emotional investment while still answering the\nwhole turn. Low effort suggests a simpler response; multiple thoughtful messages\ncan support a richer one. Do not punish short replies or overreward long ones.\n\n### Cognitive load\n\nUse one clear center of gravity. Prefer one answerable opening over stacked\nquestions, multiple decisions, or an elaborate setup. Easy to answer does not\nmean generic: a constrained, specific question is often both easier and more\ndistinctive.\n\n### Curiosity and information gaps\n\nCreate curiosity through a real specific detail, contrast, callback, or partial\nthought. Do not use clickbait, false suspense, withholding, jealousy, or a claim\nthat requires invented facts.\n\n### Self-disclosure\n\nReciprocal self-disclosure can build familiarity when it is true, relevant, and\nproportionate. Share enough to create exchange rather than interrogate, but do\nnot overshare to manufacture intimacy. Never infer a personal story from a\nphoto, interest, profession, personality label, or profile aesthetic.\n\n### Humor and play\n\nHumor works best from a shared premise. Sarcasm, dark humor, absurdity, ASCII,\nor Unicode novelty may fit the user's voice when the task and current tone\nsupport them. Tease situations, contradictions, or shared material—not\ninsecurities, appearance, identity, trauma, competence, or boundaries.\n\n### Emotional attunement\n\nRecognize meaningful emotion without diagnosing or becoming a therapist. Match\nthe weight of the disclosure: avoid jokes that bypass it, generic reassurance,\npremature advice, or exaggerated intimacy.\n\n### Topic momentum\n\nContinue when there is unanswered substance, reciprocal curiosity, a story,\nplayful tension, logistics, or a strong callback. Change topic only after the\ncurrent turn is sufficiently received. When both are needed, bridge rather than\nabruptly switching.\n\n## User-specific limits\n\n- Never propose coffee, a date, meeting, or “going out” on the user's behalf.\n- CTA may advance only to one optional Instagram or Facebook exchange unless\n  the current user explicitly changes this rule.\n- Do not use generic or scenario-based questions as a default.\n- Do not reveal Wiki scoring, attractiveness, blockers, private notes,\n  personality hypotheses, or internal effort level.\n- A checklist question may be integrated only when the router selected it and\n  the current thread can carry it naturally.\n- Do not lie about the user's view to conceal a negative assessment. Use a\n  truthful neutral explanation, defer, or keep the private rubric undisclosed.\n\n## Effort calibration\n\n- `minimal`: choose one direct, low-complexity move; no deep callbacks or extra\n  discovery unless necessary.\n- `normal`: use the current thread, one specific contribution, and an organic\n  opening when useful.\n- `high`: compare at least two materially different directions, check hidden\n  callbacks and all active topics, then choose the strongest grounded option.\n- `pause`: return no message strategy until the router's question is answered.\n\nHigh effort means better analysis, not longer or more ornate output. Minimal\neffort never permits low-quality, disrespectful, or untruthful wording.\n\n## Strategy output\n\nReturn a strategy brief:\n\n- `task`, `language`, `ruleset`, and `effort`;\n- current conversational state;\n- primary and optional secondary move;\n- must-address and optional topics;\n- intended emotional tone and pressure level;\n- response architecture;\n- specific Match hook or callback, if any;\n- permitted verified self-disclosure, if any;\n- optional checklist discovery goal;\n- question form or non-question opening;\n- facts and private assessments that must remain hidden;\n- candidate directions that are materially distinct;\n- risks, assumptions, and rejection criteria.\n\nReject any direction that could be sent unchanged to almost anyone when a\nspecific grounded alternative exists.\n"
}