Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add aristoteleo/PantheonOS --skill developmental_gene_panelgit clone --depth 1 https://github.com/aristoteleo/PantheonOSWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/aristoteleo/pantheonos/developmental_gene_panel)<a href="https://agentmods.dev/skills/aristoteleo/pantheonos/developmental_gene_panel"><img src="https://agentmods.dev/badge/skills/aristoteleo/pantheonos/developmental_gene_panel/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/aristoteleo/pantheonos/developmental_gene_panel"><img src="https://agentmods.dev/badge/skills/aristoteleo/pantheonos/developmental_gene_panel.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00129 | $0.06500 |
| Opus 5 | $0.00064 | $0.03250 |
| Sonnet 5 | $0.00026 | $0.01300 |
| Haiku 4.5 | $0.00013 | $0.00650 |
Grade A, and why
Developmental Gene Panel Design Workflow scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 11d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 415 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Developmental Gene Panel Design
Use this skill INSTEAD of gene_panel_selection when the system is developmental / dynamic:
embryonic organs, in-vitro differentiation, regeneration, or any design where the user cares about
lineage origin and regulators, not only terminal cell-type classification.
Why a separate workflow (the motivating evidence)
In a mouse embryonic-heart case study, a known cardiac developmental co-repressor (a CHD-associated gene with published human evidence) was ranked by six unbiased methods on both a late (E10.5–E14.5) and an early (E6.5–E8.5) reference:
| ranking method | early reference | late (target) reference |
|---|---|---|
| cell-type DE (Wilcoxon) | top 0.6% | top 3.0% |
| trajectory pseudotime (DPT, Pearson) | top 1.1% | top 79% |
| trajectory pseudotime (DPT, Spearman) | top 1.8% | — |
| cross-stage ANOVA (within lineage) | top 1.6% | top 18.6% |
| global marker rank | — | top 21.6% |
Every method recovered the gene from early data and failed on the target-stage data. The reason is biological: the gene rises ~19-fold along the differentiation trajectory, then becomes broadly expressed and static once cells are differentiated. Literature lookup did not rescue it either — theme-level queries ("cardiac transcription factors") return only canonical genes.
Conclusion driving this skill: developmental regulators must be recovered from an EARLY reference by TRAJECTORY-AWARE ranking. Neither target-stage DE nor literature search is sufficient.
Workflow enforcement
Steps 1–6 are mandatory and ordered. Do not silently skip a step; if a step is impossible (e.g. no early reference exists), say so explicitly in the report and justify the fallback.
Workdir / agents / reporting
Work in the workdir given by the leader. If sub-agents exist, delegate execution to analysis_expert
and interpretation to biologist; otherwise run the analysis yourself with the notebook/python toolset.
Write report_analysis.md (Summary / Data / Methods / Results / Key findings / Limitations).
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 11d ago First seen · 415 lines · 129 tokens per session scan A e82e432d4219
Developmental Gene Panel Design Workflow is a skill published in the GitHub repository aristoteleo/PantheonOS (485 stars, last pushed today), licensed BSD-2-Clause. It adds 129 tokens to every session and 6,500 once invoked, about $0.0006 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
sc-pseudotime
Load when ordering cells along a developmental trajectory in a normalised scRNA AnnData via DPT, Palantir, VIA, CellRank, Slingshot (R), or Monocle3 (R). Skip when ranking marker genes per cluster (use sc-markers); RNA velocity vector fields (use sc-velocity).
spatial-trajectory
Load when inferring pseudotime / lineage trajectories on a preprocessed spatial AnnData via DPT (default — diffusion pseudotime), CellRank (terminal-state + fate-probability), or Palantir (waypoint branch probabilities). Skip when the data has spliced/unspliced layers and you want velocity-driven dynamics (use…
sc-consensus-pseudotime
Load when you want a single-cell pseudotime ordering robust to the choice of trajectory method — fanning out DPT/Palantir/VIA from a shared root, rank-aligning them, and voting a consensus pseudotime with per-cell uncertainty. Skip when you have branching multi-lineage trajectories; no defined root.
trajectory-analysis
Single-cell trajectory inference pipeline covering diffusion pseudotime (DPT), PAGA, RNA velocity with scVelo, and fate mapping with CellRank. Use when the user mentions pseudotime, trajectory, lineage, differentiation, RNA velocity, scVelo, CellRank, PAGA, diffusion map, fate probabilities, terminal states…
spatial-trajectory
Trajectory inference and pseudotime analysis for spatial transcriptomics data.
spatial-transcriptomics-spatial-data-io
Load spatial transcriptomics data from Visium, Xenium, MERFISH, Slide-seq, and other platforms using Squidpy and SpatialData. Use this skill when: (1) Loading Visium spatial transcriptomics data from Space Ranger output, (2) Loading Xenium single-cell resolution spatial data, (3) Loading MERFISH, CosMx, or other…