Hermes Tweet

Use this ASK-compatible wrapper when a Hermes Agent user needs the native Hermes
Tweet plugin for X/Twitter automation through Xquik.

Install

Install the native plugin in Hermes Agent:

hermes plugins install Xquik-dev/hermes-tweet --enable
hermes tools list

Set XQUIK_API_KEY in the Hermes runtime environment before using authenticated
read or action tools. Do not paste the key into chat.

When to Use

Use Hermes Tweet for:

• scrape/search tweets or search Twitter/X
• read tweet replies and tweet details
• look up users and public profiles
• monitor tweets or accounts
• export followers and following lists
• post tweets/replies, send DMs, or automate X actions after explicit approval

Tool Flow

1. Use tweet_explore to find the catalog endpoint.
2. Use tweet_read for public read-only endpoints.
3. Use tweet_action only for approved writes, private reads, monitors,
   webhooks, extraction jobs, giveaway draws, or media operations.

Safety

• Never ask for API keys, passwords, cookies, or TOTP secrets.
• Never pass credentials in tool arguments.
• Use only catalog-listed /api/v1/... endpoints.
• Copied endpoint URLs are accepted only when they resolve to catalog-listed paths.
• Keep write actions gated behind HERMES_TWEET_ENABLE_ACTIONS=true.
• Summarize the exact action before posting, replying, sending DMs, or changing
  account state.

Permissions and Trust

• Tool scope: use only tweet_explore, tweet_read, and tweet_action through
  the enabled Hermes Tweet toolset.
• Network scope: call only catalog-listed Xquik API routes through those tools.
  Do not create direct HTTP fallbacks.
• File scope: do not write files, logs, screenshots, cached payloads, or
  credentials unless the user asks for an explicit export workflow.
• Environment scope: check only whether XQUIK_API_KEY,
  HERMES_TWEET_ENABLE_ACTIONS, and HERMES_ENABLE_PROJECT_PLUGINS are
  configured. Never request or echo values.
• Output: return concise Markdown summaries, action previews, or JSON-like tool
  payloads. tweet_action may change account or workflow state only after
  explicit approval.
• Release gate: do not present this skill as NVIDIA-verified unless the release
  includes a clean SkillSpector review, skill-card.md, Tier-3 eval data,
  BENCHMARK.md, skill.oms.sig, and signature verification instructions.