How do I install the Visit Jumbo Deep skill?

Run npx skillmds@latest add bbrysonelite-max/visit-jumbo-deep in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Visit Jumbo Deep safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Visit Jumbo Deep?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Visit Jumbo Deep free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Visit Jumbo Deep?

bbrysonelite-max (@bbrysonelite-max) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Visit Jumbo Deep by bbrysonelite-max

Name: Visit Jumbo Deep
Author: bbrysonelite-max

Visit Jumbo Deep

Deep grounding dossier — chains /visit-jumbo's quick brief first, then drills per-lane Jumbo queries, repo/PR states, and a full health verdict into a working dossier. Use when Brent says "visit jumbo deep", "deep grounding", "full dossier", "ground yourself deep", after a long gap between sessions, or before taking on a consequential build. Slower (several minutes). For a fast session start use /visit-jumbo.

File contents

visit-jumbo-deep

CHAIN: Stage 1 IS the /visit-jumbo skill — execute it in full (health gate → layers → one-screen brief). Then extend. Same laws: Jumbo beats guessing — never guess or assume what Jumbo can answer (Brent 2026-08-18); every line stamped; live re-ground before acting where a live query exists.

Stage 2 — drill each ALIVE lane (from Stage 1's brief)

For each lane (cap 6):

memory_search the lane by name + "decision OR owed OR blocked" (limit 5).
If the lane has a repo (MEMORY.md names it): git -C <repo> log --oneline -3
- gh pr list --repo <owner/repo> --limit 5 — real PR states, not remembered ones.
One paragraph per lane: state, last receipt (UTC), open items, contradictions between memory and repo (repo wins; flag the drift).

Stage 3 — system truth floor

Full /jumbo-health pass (its 4-step verdict block, receipts).
Ground-truth cards: if a Tiger lane is alive, run bash ~/.claude/hooks/ground-truth-card.sh --refresh and quote the fresh card — never cite its numbers from memory.
Scheduled-jobs pulse: latest lines of ~/Library/Logs/memory-check.log and memory-sweep.log; PROPOSALS.md pending-row count if a sweep has run.

Stage 4 — the dossier

DOSSIER <utc>  (jumbo GO|RED, cards fresh <utc>)
WHO + HOW      2-3 lines
LANES          per-lane paragraphs from Stage 2 (receipts inline)
DRIFT          every memory-vs-repo contradiction found (or "none found")
OWED           consolidated, dated, UNVERIFIED-labeled where unproven
RISKS          top 2-3 from tonight's state (grounded only — no generic filler)

Close with exactly one line: Most likely anchor: <X> — confirm or redirect.

Boundaries

Everything read-only; no captures/supersedes/repairs/deploys.
If Stage 1's health gate fails, STOP there — a deep dossier from a sick store is confident garbage.
Claude Code sessions only (browser agents wait on the scoped-token build).

---name: visit-jumbo-deepdescription: Deep grounding dossier — chains /visit-jumbo's quick brief first, then drills per-lane Jumbo queries, repo/PR states, and a full health verdict into a working dossier. Use when Brent says "visit jumbo deep", "deep grounding", "full dossier", "ground yourself deep", after a long gap between sessions, or before taking on a consequential build. Slower (several minutes). For a fast session start use /visit-jumbo.---# visit-jumbo-deepCHAIN: Stage 1 IS the `/visit-jumbo` skill — execute it in full (health gate →layers → one-screen brief). Then extend. Same laws: Jumbo beats guessing —never guess or assume what Jumbo can answer (Brent 2026-08-18); every linestamped; live re-ground before acting where a live query exists.## Stage 2 — drill each ALIVE lane (from Stage 1's brief)For each lane (cap 6):1. `memory_search` the lane by name + "decision OR owed OR blocked" (limit 5).2. If the lane has a repo (MEMORY.md names it): `git -C <repo> log --oneline -3` + `gh pr list --repo <owner/repo> --limit 5` — real PR states, not remembered ones.3. One paragraph per lane: state, last receipt (UTC), open items, contradictions between memory and repo (repo wins; flag the drift).## Stage 3 — system truth floor1. Full `/jumbo-health` pass (its 4-step verdict block, receipts).2. Ground-truth cards: if a Tiger lane is alive, run `bash ~/.claude/hooks/ground-truth-card.sh --refresh` and quote the fresh card — never cite its numbers from memory.3. Scheduled-jobs pulse: latest lines of `~/Library/Logs/memory-check.log` and `memory-sweep.log`; PROPOSALS.md pending-row count if a sweep has run.## Stage 4 — the dossier```DOSSIER <utc> (jumbo GO|RED, cards fresh <utc>)WHO + HOW 2-3 linesLANES per-lane paragraphs from Stage 2 (receipts inline)DRIFT every memory-vs-repo contradiction found (or "none found")OWED consolidated, dated, UNVERIFIED-labeled where unprovenRISKS top 2-3 from tonight's state (grounded only — no generic filler)```Close with exactly one line:`Most likely anchor: <X> — confirm or redirect.`## Boundaries- Everything read-only; no captures/supersedes/repairs/deploys.- If Stage 1's health gate fails, STOP there — a deep dossier from a sick store is confident garbage.- Claude Code sessions only (browser agents wait on the scoped-token build).

Frequently asked questions