wjlgatech · wjlgatech · Jul 19, 2026 · Jul 19, 2026
diff --git a/Makefile b/Makefile
@@ -1,4 +1,4 @@
-.PHONY: build validate audit eval site graph track synthesize recap-dry clockbench fair repos check help
+.PHONY: build validate audit eval site graph track synthesize recap-dry clockbench fair repos hub check help
 
 help:
 	@echo "longevity-loop — an AI-native compounding loop for aging science"
@@ -10,7 +10,8 @@ help:
 	@echo "  make clockbench  Cross-clock disagreement benchmark (measures gaps-analysis G1)"
 	@echo "  make fair        Score open aging datasets for FAIR/reproducibility → docs/FAIR.md (G2)"
 	@echo "  make repos       Per-repo KG + agentic-tooling generation recipes → docs/REPOS.md (hub)"
-	@echo "  make check       validate + audit + eval + clockbench + fair + repos + build drift-gate (CI finish line)"
+	@echo "  make hub         Validate the JIT deep artifacts under hub/ (add --refresh for staleness)"
+	@echo "  make check       validate + audit + eval + clockbench + fair + repos + hub + build drift-gate (CI finish line)"
 
 build:
 	python3 scripts/build.py
@@ -59,13 +60,20 @@ fair:
 repos:
 	python3 scripts/repos.py
 
-# Finish line: data well-formed, still an AI-native loop, executions honestly eval'd
-# (no fake ✅), the clock-disagreement + FAIR + repo-recipe artifacts self-verify, and docs match.
+# Validate the hub's JIT deep artifacts (hub/<repo>/: KG + skill + SHA-stamped manifest).
+# `make hub` = offline integrity gate; `python3 scripts/hub_gen.py --refresh` = network staleness.
+hub:
+	python3 scripts/hub_gen.py --check
+
+# Finish line: data well-formed, still an AI-native loop, executions honestly eval'd (no fake ✅),
+# the clock-disagreement + FAIR + repo-recipe + hub artifacts self-verify, and docs match.
 check: validate audit eval
 	python3 scripts/clockbench.py --selftest
 	python3 scripts/fair.py --selftest
 	python3 scripts/fair.py --gate
 	python3 scripts/fair.py --check
 	python3 scripts/repos.py --selftest
 	python3 scripts/repos.py --check
+	python3 scripts/hub_gen.py --selftest
+	python3 scripts/hub_gen.py --check
 	python3 scripts/build.py --check
diff --git a/README.md b/README.md
@@ -157,7 +157,7 @@ Sequenced bridge: convert the computational finding + your credential network (G
 
 <h2 id="execution">✅ Roadmap Execution</h2>
 
-**12/23 done (52%).** Each execution is a checkbox with a before→after eval; a box only ticks with a real result (`done` requires a non-pending before AND after — no evidence ⇒ not done, gated in CI).
+**13/24 done (54%).** Each execution is a checkbox with a before→after eval; a box only ticks with a real result (`done` requires a non-pending before AND after — no evidence ⇒ not done, gated in CI).
 
 - [x] **E1** (P1 · Foundation) — Stand up the AI-native loop (repo + self-audit + gated `make check`)
 - [x] **E2** (P1 · Foundation) — Map the field into a verified knowledge base
@@ -182,6 +182,7 @@ Sequenced bridge: convert the computational finding + your credential network (G
 - [x] **E21** (P1 · Foundation) — Problem map: top aging problems as a relationship graph + the one target + low-hanging fruit
 - [x] **E22** (P1 · Foundation) — Publish the tracker.json data contract for the interactive portfolio instance (data-first slice)
 - [x] **E23** (P1 · Foundation) — Hub architecture (knowledge·tooling·experts per cited repo) + the per-repo generation-recipe layer
+- [x] **E24** (P1 · Foundation) — Hub Phase-2: JIT deep artifacts for one repo (pyaging) — knowledge graph + gated skill, SHA-stamped
 
 ### Eval reports — before → after
 
@@ -210,6 +211,7 @@ Sequenced bridge: convert the computational finding + your credential network (G
 | ✅ E21 | a clear, visual answer to 'what are we shooting, and what's the quick win' | gaps ranked (gaps-analysis.md) but relationships + the single target + quick wins were implicit | docs/PROBLEMS.md — Mermaid relationship graph (G1 measurement = root bottleneck we target) + top-problems table + how/when/why + 3 low-hanging fruit with done-when; README pointer |
 | ✅ E22 | a stable, derived JSON contract the agentic-portfolio instance can consume (no drift) | no machine-readable mission/progress feed for an external interactive surface | site/tracker.json — mission + 48% progress + 21 milestones + roadmap + low-hanging + frontier + 12 hubs + links, generated by build_site.py, deployed on Pages |
 | ✅ E23 | a fresh, cheap, future-proof way to offer a KG + agentic tooling for each cited repo | cited repos listed in stack.yml but no KG/tooling per repo; no design for how to do it well | docs/HUB_ARCHITECTURE.md (index+JIT design; reuse graphify/understand/reverse; MCP; SHA-cached) + scripts/repos.py → docs/REPOS.md (22 repos, per-repo KG + reverse recipes), gated in make check |
+| ✅ E24 | a real, grounded, provenance-stamped KG + skill generated from a cited repo, with a freshness gate | hub had the recipe layer only; no deep artifact proving the JIT layer works | hub/pyaging/ — knowledge-graph.json (20 nodes/edges) + compute-aging-clocks SKILL.md + manifest (SHA c2b3000e3c2c); scripts/hub_gen.py gates integrity (make check) + checks live staleness (FRESH); grounded in pyaging's real API |
 
 ---
 
@@ -232,7 +234,7 @@ Sequenced bridge: convert the computational finding + your credential network (G
 
 <h2 id="hub">🧩 Knowledge · Tooling · Experts hub</h2>
 
-This repo is a **hub**: for each cited repo it offers a knowledge graph (knowledge) + agentic tooling (skills/plugins/workflows), and links them to the experts who build them. Design (high-quality · fast · cheap · fresh · future-proof): **[docs/HUB_ARCHITECTURE.md](docs/HUB_ARCHITECTURE.md)** · per-repo generation recipes: **[docs/REPOS.md](docs/REPOS.md)**. Fresh index, just-in-time depth — deep artifacts are generated on demand and cached by commit SHA, never committed stale.
+This repo is a **hub**: for each cited repo it offers a knowledge graph (knowledge) + agentic tooling (skills/plugins/workflows), and links them to the experts who build them. Design (high-quality · fast · cheap · fresh · future-proof): **[docs/HUB_ARCHITECTURE.md](docs/HUB_ARCHITECTURE.md)** · per-repo generation recipes: **[docs/REPOS.md](docs/REPOS.md)**. Fresh index, just-in-time depth — deep artifacts are generated on demand and cached by commit SHA, never committed stale. First deep exemplar (Phase 2): **[hub/pyaging](hub/pyaging)** — a grounded knowledge graph + a gated skill, SHA-stamped, with a live staleness check (`make hub` / `scripts/hub_gen.py --refresh`).
 
 ---
 

diff --git a/data/executions.yml b/data/executions.yml
@@ -168,3 +168,12 @@
     before: "cited repos listed in stack.yml but no KG/tooling per repo; no design for how to do it well"
     after: "docs/HUB_ARCHITECTURE.md (index+JIT design; reuse graphify/understand/reverse; MCP; SHA-cached) + scripts/repos.py → docs/REPOS.md (22 repos, per-repo KG + reverse recipes), gated in make check"
   proof: "docs/HUB_ARCHITECTURE.md · scripts/repos.py · docs/REPOS.md · make check"
+- id: E24
+  phase: "P1 · Foundation"
+  item: "Hub Phase-2: JIT deep artifacts for one repo (pyaging) — knowledge graph + gated skill, SHA-stamped"
+  done: true
+  eval:
+    metric: "a real, grounded, provenance-stamped KG + skill generated from a cited repo, with a freshness gate"
+    before: "hub had the recipe layer only; no deep artifact proving the JIT layer works"
+    after: "hub/pyaging/ — knowledge-graph.json (20 nodes/edges) + compute-aging-clocks SKILL.md + manifest (SHA c2b3000e3c2c); scripts/hub_gen.py gates integrity (make check) + checks live staleness (FRESH); grounded in pyaging's real API"
+  proof: "hub/pyaging/* · scripts/hub_gen.py · docs/HUB_ARCHITECTURE.md"
diff --git a/hub/pyaging/SKILL.md b/hub/pyaging/SKILL.md
@@ -0,0 +1,56 @@
+---
+name: compute-aging-clocks
+description: Compute a panel of biological-age (epigenetic) clocks from methylation/omics data with pyaging, then hand the panel to clockbench to quantify cross-clock disagreement. Trigger when you need biological-age estimates, a clock panel, or the disagreement signal for the Biomarkers-of-Aging / Turn-01 work.
+source: rsinghlab/pyaging @ c2b3000e3c2c
+license: BSD-3-Clause (pyaging)
+generated_by: longevity-loop hub (scripts/hub_gen.py) — grounded in the pyaging repo + docs, not invented
+---
+
+# compute-aging-clocks (via pyaging)
+
+Generated by the longevity-loop **hub** (docs/HUB_ARCHITECTURE.md) as the Phase-2 exemplar of
+just-in-time deep tooling. Grounded in pyaging's real API at the pinned `source` SHA — regenerate
+when that SHA changes (`scripts/hub_gen.py --refresh` flags staleness).
+
+## When to use
+You have methylation (or RNA-seq / ATAC / blood-chemistry) data and need biological-age estimates
+from a **panel** of clocks — e.g. to test whether cross-clock *disagreement* adds signal
+(gaps-analysis G1, Turn 01), or to score biological age for the Biolearn leaderboard.
+
+## Install
+```bash
+pip install pyaging   # PyTorch-based; GPU optional but faster
+```
+
+## The workflow (pyaging's real API)
+```python
+import pyaging as pya
+import pandas as pd
+
+# 1. Load your data as a DataFrame (rows = samples, cols = features/CpGs), then -> AnnData.
+#    (EPICv2 arrays: aggregate probes first.)
+# df = pya.pp.epicv2_probe_aggregation(df)          # only for EPICv2
+adata = pya.pp.df_to_adata(df, metadata_cols=["gender", "tissue_type", "dataset"],
+                           imputer_strategy="knn")
+
+# 2. Predict a PANEL of clocks in one call (170+ available; these are real clock names).
+pya.pred.predict_age(adata, ["Horvath2013", "AltumAge", "PCGrimAge", "GrimAge2", "DunedinPACE"])
+
+# 3. Read results — one column per clock in adata.obs; clock metadata in adata.uns.
+print(adata.obs.head())          # predicted ages per sample, per clock
+# adata.uns["Horvath2013_metadata"], adata.uns["<clock>_missing_features"]
+```
+
+## Hand off to the loop (why this skill exists)
+Export the per-clock predicted ages and feed them to this repo's benchmark to get the
+**cross-clock disagreement** number (the measurable core of G1):
+```bash
+# adata.obs[clock_cols] -> a CSV (rows = samples, header = clock names), then:
+python3 scripts/clockbench.py --input clocks.csv
+```
+That closes the loop: pyaging (compute the panel) → clockbench (measure disagreement) → Turn 01.
+
+## Discipline
+Predicted ages are **computational estimates on your data** — label them as such and keep them
+separate from any wet-lab/clinical claim. Verify the clock names against the current pyaging
+ModelGallery (they evolve); `no evidence ⇒ no claim`.
diff --git a/hub/pyaging/knowledge-graph.json b/hub/pyaging/knowledge-graph.json
@@ -0,0 +1,54 @@
+{
+  "schema": "longevity-loop hub knowledge graph — portable nodes/edges JSON",
+  "source": {
+    "repo": "rsinghlab/pyaging",
+    "url": "https://github.com/rsinghlab/pyaging",
+    "sha": "c2b3000e3c2c",
+    "grounded_in": "repo README + pyaging.readthedocs.io tutorials (verbatim API)"
+  },
+  "nodes": [
+    {"id": "repo:pyaging", "label": "pyaging", "type": "repo", "note": "170+ GPU-optimized aging clocks; BSD-3"},
+    {"id": "mod:pp", "label": "pya.pp", "type": "module", "note": "preprocessing / data prep"},
+    {"id": "mod:pred", "label": "pya.pred", "type": "module", "note": "clock prediction"},
+    {"id": "mod:data", "label": "pya.data", "type": "module", "note": "example data / loaders"},
+    {"id": "fn:df_to_adata", "label": "pya.pp.df_to_adata", "type": "function", "note": "DataFrame -> AnnData (metadata_cols, imputer_strategy='knn')"},
+    {"id": "fn:epicv2", "label": "pya.pp.epicv2_probe_aggregation", "type": "function", "note": "EPICv2 probe aggregation"},
+    {"id": "fn:predict_age", "label": "pya.pred.predict_age", "type": "function", "note": "predict_age(adata, [clock_names]) -> results in adata.obs"},
+    {"id": "concept:anndata", "label": "AnnData", "type": "concept", "note": "the core data container"},
+    {"id": "concept:clock", "label": "aging clock (170+)", "type": "concept"},
+    {"id": "clock:Horvath2013", "label": "Horvath2013", "type": "clock"},
+    {"id": "clock:AltumAge", "label": "AltumAge", "type": "clock"},
+    {"id": "clock:PCGrimAge", "label": "PCGrimAge", "type": "clock"},
+    {"id": "clock:GrimAge2", "label": "GrimAge2", "type": "clock"},
+    {"id": "clock:DunedinPACE", "label": "DunedinPACE", "type": "clock"},
+    {"id": "modality:dnam", "label": "DNA methylation", "type": "modality"},
+    {"id": "modality:rna", "label": "RNA-seq", "type": "modality"},
+    {"id": "modality:atac", "label": "ATAC-seq", "type": "modality"},
+    {"id": "modality:histone", "label": "histone marks", "type": "modality"},
+    {"id": "modality:chem", "label": "blood chemistry", "type": "modality"},
+    {"id": "ext:clockbench", "label": "longevity-loop clockbench (G1)", "type": "downstream", "note": "consumes the clock panel to measure cross-clock disagreement"}
+  ],
+  "edges": [
+    {"source": "repo:pyaging", "target": "mod:pp", "rel": "contains"},
+    {"source": "repo:pyaging", "target": "mod:pred", "rel": "contains"},
+    {"source": "repo:pyaging", "target": "mod:data", "rel": "contains"},
+    {"source": "mod:pp", "target": "fn:df_to_adata", "rel": "defines"},
+    {"source": "mod:pp", "target": "fn:epicv2", "rel": "defines"},
+    {"source": "mod:pred", "target": "fn:predict_age", "rel": "defines"},
+    {"source": "fn:df_to_adata", "target": "concept:anndata", "rel": "produces"},
+    {"source": "fn:predict_age", "target": "concept:anndata", "rel": "consumes"},
+    {"source": "fn:predict_age", "target": "concept:clock", "rel": "predicts"},
+    {"source": "clock:Horvath2013", "target": "concept:clock", "rel": "is-a"},
+    {"source": "clock:AltumAge", "target": "concept:clock", "rel": "is-a"},
+    {"source": "clock:PCGrimAge", "target": "concept:clock", "rel": "is-a"},
+    {"source": "clock:GrimAge2", "target": "concept:clock", "rel": "is-a"},
+    {"source": "clock:DunedinPACE", "target": "concept:clock", "rel": "is-a"},
+    {"source": "repo:pyaging", "target": "modality:dnam", "rel": "supports"},
+    {"source": "repo:pyaging", "target": "modality:rna", "rel": "supports"},
+    {"source": "repo:pyaging", "target": "modality:atac", "rel": "supports"},
+    {"source": "repo:pyaging", "target": "modality:histone", "rel": "supports"},
+    {"source": "repo:pyaging", "target": "modality:chem", "rel": "supports"},
+    {"source": "fn:predict_age", "target": "ext:clockbench", "rel": "feeds"}
+  ],
+  "counts": {"nodes": 20, "edges": 20}
+}
diff --git a/hub/pyaging/manifest.json b/hub/pyaging/manifest.json
@@ -0,0 +1,16 @@
+{
+  "repo": "rsinghlab/pyaging",
+  "url": "https://github.com/rsinghlab/pyaging",
+  "source_sha": "c2b3000e3c2c",
+  "generated_at": "2026-07-19",
+  "license": "BSD-3-Clause",
+  "stars_at_generation": 127,
+  "one_line": "A Python compendium of GPU-optimized biological aging clocks (170+ published clocks).",
+  "artifacts": ["knowledge-graph.json", "SKILL.md"],
+  "engines": {
+    "knowledge_graph": "grounded from the pyaging repo + readthedocs (production: graphify / understand-anything)",
+    "skill": "grounded from the pyaging real API (production: anyagent reverse -> refine)"
+  },
+  "provenance_note": "Every fact is grounded in the pyaging repo/docs at source_sha. This is the Phase-2 exemplar of the JIT deep layer (docs/HUB_ARCHITECTURE.md): regenerate when source_sha changes — `scripts/hub_gen.py --refresh` flags staleness. No evidence => no claim.",
+  "why": "pyaging computes the epigenetic-clock panel that feeds scripts/clockbench.py (gaps-analysis G1) and Turn 01 — the loop's #1 measurement target."
+}
diff --git a/scripts/build.py b/scripts/build.py
@@ -150,7 +150,9 @@ def render_readme() -> str:
           "Design (high-quality · fast · cheap · fresh · future-proof): "
           "**[docs/HUB_ARCHITECTURE.md](docs/HUB_ARCHITECTURE.md)** · per-repo generation recipes: "
           "**[docs/REPOS.md](docs/REPOS.md)**. Fresh index, just-in-time depth — deep artifacts are "
-          "generated on demand and cached by commit SHA, never committed stale.", "", "---", ""]
+          "generated on demand and cached by commit SHA, never committed stale. First deep exemplar "
+          "(Phase 2): **[hub/pyaging](hub/pyaging)** — a grounded knowledge graph + a gated skill, "
+          "SHA-stamped, with a live staleness check (`make hub` / `scripts/hub_gen.py --refresh`).", "", "---", ""]
 
     # What we're shooting (static pointer to the problem map)
     L += ['<h2 id="target">🎯 What we\'re shooting</h2>', "",