Each Playbook teaches one pattern perea has run in production — names the engine it came from, cites the production evidence that validates it, and is sized for an agent to read in a single fetch. Audience: founders and agents building a similar AI-content pipeline.
The /loop + fat-tick + state-machine + lock-file + kill-switch shape that lets a Claude Code skill run a production content pipeline unattended — and the sibling-independence rule that prevents engines from corrupting each other.
perea-research-engine· ~2,500 words· Public draft· v1.0An optimizer-written long-form post structured around 5–7 Quotables as cited evidence — sized for X's 25k Premium ceiling, shaped for the algorithm's dwell-time signal.
tick-X-post-from-research· ~1,800 words· Public draft· v1.0Free tone, locked numerics, locked named entities, no forward-looking predictions — the brand contract that lets an algorithm-optimizer rewrite your content without drifting from the source paper.
tick-X-post-from-research· ~2,300 words· Public draft· v1.0A three-tier drift-mitigation regex/phrase list that rejects engagement-bait, urgency manipulation, and forward-looking confidence before the post-hoc reviewer ever runs.
tick-X-post-from-research· ~1,700 words· Public draft· v1.0A literal-substring faithfulness gate that turns a research paper into a reusable set of cited atomic claims — readable in one fetch, consumable by every downstream channel.
perea-research-engine· ~2,400 words· Public draft· v1.0Skip the calibration window. Schedule from day one — to a queue 14 days ahead, two posts per day, at two fixed UTC slots — and let the queue itself act as your QA gate.
tick-X-post-from-research· ~1,700 words· Public draft· v1.0