Twitter Agent

1. What Twitter Agent Does

Twitter Agent answers quantum analysis claims about social attention and sentiment around a subnet on X/Twitter, mention volume, engagement, account reach, and sentiment split, using evidence it has actually collected rather than the model's own memory.

Sentiment scores themselves are computed offline by a dedicated classifier, never invented by the quantum analysis turn model, so a verdict reasons over retrieved metrics rather than guessing at tone.

2. Where It Sits in the Tao Status Economy

Twitter Agent is a consumer of miner-contributed LLM keys, preferring a pooled miner key for the required provider and model and falling back to its own static key, invisible to scoring, only when contribution is disabled, the pool declines, or the vendor is unsupported. One lease, one LLM call, one report per turn, with no retry loop.

3. The Turn Pipeline

• No mentions in the claim's window: the agent declines and reports nothing. A success report here would wrongly credit a miner for a call that never happened, so none is filed.
• An exception during the LLM call is always reported as a real key outcome, success false, with the exception's category.
• Reporting is best-effort: a reporting failure is logged and swallowed rather than failing the turn, and fallback keys with no hotkey are never reported.

4. Quality Grading

5. Reporting

One report is filed per turn, feeding directly into the reporting miner's scoring window:

{
  "hotkey": "5F...",
  "success": true,
  "latency_ms": 1655.8,
  "error_category": null,
  "quality_score": 0.75,
  "lease_token": "the token from this turn's acquire grant"
}

• success means no exception escaped the turn; on failure, error_category carries the exception's class name.
• latency_ms spans the whole agent run: retrieval and the LLM call.
• lease_token is required, and each token is single-use.

6. Simulated Example

A quantum analysis turn on the claim: "Subnet 5 saw a sentiment shift after its release announcement."

1. 1

   Lease

   A pooled miner key is acquired for the session's provider and model.

2. 2

   Retrieve

   The claim resolves to a seven-day window around the announcement date. Retrieval returns 118 mentions and 9,400 combined engagement, with positive sentiment rising from 54% before the window to 71% after, plus three high-engagement posts referencing the release. 118 mentions clears the medium-confidence threshold but not the high one.

3. 3

   One LLM call

   The model returns a SUPPORTED verdict citing the pre- and post-window sentiment split and mention count, with a caveat about the modest sample size.

4. 4

   Report

   The turn is graded 0.75, settled at medium confidence, and reported.

5. 5

   Variant: no mentions

   A subnet with no tracked mentions in the window causes the agent to decline; the leased LLM is never called and no report is filed.

6. 6

   Variant: provider failure

   A raised error is reported as success false with the exception's category; the miner's window absorbs one failure and the key goes on cooldown.

7. Operating Limits