{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreiatic4sxnlogda6pogf3cmyrtvadc2g6miz5hmk3hwnmofktdvkdm",
    "uri": "at://did:plc:lk3jfj3zq4k4wxnk474axylu/app.bsky.feed.post/3me6ttbx3jg52"
  },
  "path": "/t/testing-a-sharp-tongued-ai-persona-looking-for-prompt-tweaks/1369265?page=9#post_171",
  "publishedAt": "2026-02-06T11:00:54.000Z",
  "site": "https://community.openai.com",
  "textContent": "Thanks for the feedback.\n\nElectric-Shy is a full SKILL.md specification rather than a simple persona prompt. It’s designed for cross-platform stability and has been tested extensively on both Claude and ChatGPT.\n\nRegarding “trigger words”, in this skill they refer to the specific activation boundary cues defined inside the SKILL.md file (Electric-Shy persona/ Electric-Shy ON/ Use Electric-Shy),not improvised stabilizing prompts.\n\nIf those internal triggers aren’t used (or if the skill file isn’t fully loaded), the system may fall back to default behavior, which can look like instability. When the skill is loaded and activated as specified, it should behave the same way for any user.",
  "title": "Testing a Sharp-Tongued AI Persona — Looking for Prompt Tweaks"
}