parker-brain

How this brain saves, syncs, and clones itself on GitHub — the required procedure for EVERY git push, pull, fetch, clone, commit, or gh command in a Parker Brain repo. Covers the credentials (the setup_parker_brain tool, never the user's own login, never gh), submodules, merge conflicts, and the rare self-hosted exception. Use whenever saving work to GitHub, syncing or updating the repo, cloning the brain, fixing a failed push or auth error, or when asked to save, back up, or share the brain.

The front door to building a brand's brain. Run this when standing up a new brand — all it needs is the Parker MCP connected. It welcomes the person, provisions the brand's own repo through the setup_parker_brain tool (Parker MCP), clones that repo and works inside it, connects and tests the data, asks the short brand intake (the things Parker genuinely can't observe), runs the full build in dependency order, saves it to GitHub, and hands off to get-started. Follows prompts/onboarding-runner.md as the detailed method. Use when someone says set up a brand, build the brain, onboard a new brand, or get started building. Also the right entry when the brand's brain already exists — it detects a built brain and retrieves it, handing off to get-started instead of rebuilding.

The weekly check that keeps this brain current with the parker-brain standard — without overriding anything the team chose. Compares the pinned parker-system/ release against the newest factory tag and offers the bump as one short question; on a yes the migrations run and a script re-syncs the copied skills deterministically. What the team declines stays declined; what they modified stays theirs. Use weekly as a scheduled routine, or when asked to update the brain, check for updates, sync with the standard, or "what's new in Parker".

One-time setup for a fresh brand brain instance. Registers the standing routines (refresh-context, dream, harvest+evaluate ideas, research-loops, update-brain, self-improve) as scheduled cloud agents so they run on cadence without being asked. The job definitions already travel with this repo; this skill arms the schedules, which are per-account and can't be committed. Run once after cloning the brain into a new Claude Code cloud instance, or when asked to set up / register / schedule the routines.

Generate AI ad assets — Veo 3 video prompts and AI image static prompts — with brand context, compliance, and forensic direction. The output is prompt text the user pastes into the AI tool, not the final asset.

Write headlines for ads — static, video text overlays, paid social, landing page hero. Picks the right structural pattern for the brand type (lifestyle vs problem/solution), the awareness stage, and the customer language available.

Write hooks for a specific ad — the first three seconds that earn the click and the watch. Uses the proven hook formats from the canonical taxonomy, picks the right one for the brand and audience, and produces hooks built around a specific ICP and emotion.

The full iteration flow in one skill — choose which ads to iterate on, then make the iterations. If the user names a specific ad, go straight to making iterations; if they ask account-level (which ads should we iterate on, what to iterate this month), select the ads first, then make iterations on each.

Write ad scripts that sound spoken, not written, by adapting proven reference ads into the brand's own voice fingerprint. AI cannot write good scripts cold — references are the default input and the brand script-voice profile is the sound. Every script passes the spoken-script-voice AI-tells check and read-aloud test. Net-new from scratch is the rare fallback.

Make a correct update to Parker's prompts, system docs, rubrics, knowledge docs, training corpus, or brand outputs — and propagate the change everywhere it needs to land. Use this skill whenever you are changing any part of the Parker context-engineering system, including tightening a rule, adding a principle, propagating a pattern, adding or modifying a prompt, adding a knowledge doc, or evolving a skill. The skill exists because Parker's docs are highly cross-referenced and inheritance-driven, and updates that look local often have non-obvious downstream impact.

Analyze Meta ad account data — read what's spending, what's converting, what's fatiguing, and where the audience is being reached. Use when the user wants an account audit, a single-ad diagnosis, a placement read, a demographic breakdown, or a fatigue check.

Maintain a brand-specific idea bank from context docs, research, Parker data, user conversations, manual ideas-tab saves, expert signals, competitor ads, organic videos, and internal Parker observations.

Turn a promoted open loop into a hypothesis — a falsifiable prediction plus the few complementary pulls that together settle it, triangulated across the ad account, reviews, and competitors as the question needs and runnable with Parker's tools. Use when advancing a Tier-1 (or activated Tier-2) loop from the open-loops roll-up into the testing queue, or when the user says advance the open loops, form a hypothesis, or plan how to test a loop.

Push-based re-ship of the factory methodology bundle into brand brains that carry the method as copied files — the legacy vendored-parker-system layout and decoupled brains that asked for a refresh. Brains on the standard layout (parker-system as a pinned submodule) do NOT use this; they update by moving the pin via their own /update-brain routine, and migrations/v1.md converts a legacy brain to that layout. Use for a factory-initiated sweep of the remaining copy-based brains.

Keep the brand brain from going stale. Reads running-notes/refresh-schedule.md and each standing doc's refresh_by frontmatter, surfaces what is overdue or due soon, walks the Phase 1→2→3 dependency spine to catch docs made stale by upstream changes even when their own date is fine, and re-runs the generating prompt for each — carrying the prior version forward and re-stamping the dates. Use weekly as a scheduled routine, or whenever asked to refresh, re-run, or check the freshness of the brand brain or any context doc.

Decouple this brain from the public parker-brain factory so the team can develop their own version of the method. Converts the read-only parker-system/ submodule into plain files the team owns and can edit, removes the read-only guardrails, and flips the update posture so /update-brain stops offering pin bumps. Use when the team says they want to customize the method itself, own their prompts and craft docs, disconnect from Parker's updates, or go independent.

The weekly research cycle — runs the open-loops pipeline end to end so the brand's strategic questions actually get answered instead of piling up. Rolls up and grades the open loops, advances the promoted ones into hypotheses, runs the validations in the testing queue, runs due re-validations, reports what was learned, and then aligns the standing docs with the confirmed findings so the brain never says the old thing after the research says otherwise. Use weekly as a scheduled routine, or when asked to run the research, advance the loops, work the testing queue, or "what have we learned".

Arms the brand brain's standing routines (refresh-context, dream, harvest+evaluate ideas, research-loops, update-brain, self-improve) as scheduled cloud agents so they run on cadence without being asked. The job definitions already travel with this repo; this skill registers the schedules, which are per-account and can't be committed. The onboarding build runs it automatically at the stamp step (build mode, no questions); run it yourself after cloning the brain into a new Claude Code cloud instance, or any time to change a cadence or turn a routine off.

Rank the brand idea bank against the approved Phase-2 strategic roadmap. Grades the whole captured pile — which priority each idea serves, the lever it pulls, how strong the evidence is, whether it trips a roads-not-taken kill — and produces a ranked shortlist by priority that ends in a plain ranked call the sprint plan sizes into a round, plus a read of what the bank is starving for. The skeptical pass. Run after harvest-ideas (weekly), when the bank grows enough to change the rank, or when the roadmap is re-approved.

Weekly agentic idea hunt for the brand. Assembles a hunt brief (roadmap, personas, live account read, tried-and-failed, user signals), runs a cold brainstorm before any feed opens, then fans out hunter lenses — customer language, organic/format with a trend lane, old-ads corpus, far-transfer wildcard — and captures each find verbatim with its spark into idea-bank/entries with full provenance and a viewable reference link. This is the capture half of Phase 3; grading is the separate evaluate-ideas skill. Use weekly as a scheduled routine, or when asked to find, harvest, or hunt ideas for the brand.

Run the dreaming system on demand for a brand and the person using it. Read the day of comms — the conversations the user had with Parker, plus any connected comms — against everything Parker has accumulated, and come back with proposals across six buckets: context updates, skill improvements, schedules to create, new ideas to research, new open loops, and the person (what Parker now understands about them and what to tee up for them). Framed as the morning suggestion. Use when Jimmy runs /dream or asks Parker what it has been thinking about, what it would suggest, or to dream on a brand.

Parker's dreaming run for the brand and the person using it — the planning half of the living layer. Reads the day's comms (the conversations the user had with Parker first, then any connected comms) against the vault and returns proposals across six buckets — context updates, skill improvements, schedules to create, new ideas to research, new open loops, and the person (what Parker now understands about them and what to tee up for them) — into dreaming/proposals/pending. Captures verbatim; proposes, never applies; promotion runs through self-improve with the human in the loop. Use daily as a scheduled routine, or when asked to "dream", brainstorm overnight, or "what were you thinking about".

Apply the self-improvement system on demand. Read the accumulated reasoning traces and update this brand's and this user's living context to reflect what Parker has learned. Use when Jimmy runs /improve-system or asks Parker to apply what it has learned, fold the traces in, or get this brand's context up to date with recent feedback.

Capture reasoning traces from user feedback, approvals, rejections, reroutes, expert learnings, and strategic decisions so Parker improves over time.

The first-run walkthrough for a brand brain — the orientation a new owner or a teammate who just cloned the repo needs so they aren't left in decision paralysis. It first reads how familiar the person is with this kind of system and calibrates to them, then covers, in plain language and paced one thing at a time, what the brain is, what was built, how the files work, how they grow it, what Parker can actually pull and do, what skills are, how to connect their other tools, what the brain does on its own, what to automate now, how to bring teammates in, and where to start. Run on a fresh brain, when someone is new to it, or whenever asked to walk me through this / what is this / where do I start / give me the tour / I'm new here.

Parker's self-improvement governance for the brand — the disposing half of the living layer. Weekly it curates captured reasoning traces and decides which dreaming proposals get promoted, human in the loop — including the doc-alignment proposals the research cycle files. It is the gate: dreaming and research propose, self-improvement disposes. The loop pipeline itself (roll-up, hypotheses, validations, re-validations) runs in /research-loops. Use weekly as a scheduled routine, or when asked to curate learnings, promote proposals, or self-improve.

Turn a knowledge source Jimmy hands Parker — an expert video, transcript, tweet, post, newsletter, podcast, article, or report — into an exact, verbatim proposal of what Parker would change, for Jimmy to approve, modify, or deny before anything is written. Use when Jimmy drops a source and wants Parker to learn from it.

Run a hypothesis's test plan, gather the evidence, and honestly resolve the loop into one of the four validation states — validated, invalidated, inconclusive, or insufficient evidence. Use when a hypothesis is approved or proceeding and it is time to test it, or when the user says validate this, run the test, close the loop, or what did the research find.