Borrowing it
Nothing to install: this file belongs to boettiger-lab/data-workflows. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/boettiger-lab/data-workflows/main/.claude/skills/armada-pipeline/SKILL.mdgit clone --depth 1 https://github.com/boettiger-lab/data-workflowsWrote 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/boettiger-lab/data-workflows/armada-pipeline)<a href="https://agentmods.dev/skills/boettiger-lab/data-workflows/armada-pipeline"><img src="https://agentmods.dev/badge/skills/boettiger-lab/data-workflows/armada-pipeline/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/boettiger-lab/data-workflows/armada-pipeline"><img src="https://agentmods.dev/badge/skills/boettiger-lab/data-workflows/armada-pipeline.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 6 findings, up to high
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- high Supply Chain · line 85 Remote code is downloaded and executed. This bypasses code review and could introduce malicious code.Fix: Avoid downloading and executing remote scripts. Use trusted packages from PyPI/npm. If remote fetch is required, verify checksums and use HTTPS.
- high Privilege Escalation · line 102 Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.Fix: Remove references to credential paths. Use environment variables or secrets managers. For docs, use placeholder paths (e.g., /path/to/config). Never load .env or token files in production code paths.
- high Supply Chain · line 126 Remote code is downloaded and executed. This bypasses code review and could introduce malicious code.Fix: Avoid downloading and executing remote scripts. Use trusted packages from PyPI/npm. If remote fetch is required, verify checksums and use HTTPS.
- medium Data Exfiltration · line 85 Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.Fix: Verify the destination URL is trusted and necessary. Remove or replace with documented APIs. Ensure no secrets, tokens, or PII are transmitted.
- medium Rogue Agent · line 95 Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.Fix: Remove any persistence mechanisms (cron jobs, startup scripts, state files). Skills should not maintain state across sessions without explicit user consent.
- medium Tool Misuse · line 177 Tool defaults are unsafe or overly permissive (e.g. disabled TLS verification, no authentication, world-writable permissions). Unsafe defaults widen the attack surface.Fix: Override unsafe defaults with secure settings (verify=True, auth required, restrictive permissions). Review and harden all tool configurations.
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 | $0.00130 | $0.05644 |
| Opus 5.5 | $0.00052 | $0.02258 |
| Sonnet 5.5 | $0.00026 | $0.01129 |
| Haiku 4.5 | $0.00013 | $0.00564 |
Grade A, and why
armada-pipeline scanned grade A with 1 finding 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 10d 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.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
curl -s -X POST "https://authentik.nrp-nautilus.io/application/o/device/" \ How it starts
The opening of the file, as written. The whole thing — 428 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Armada Pipeline (NRP)
Why Armada: microslicing, not requeueing
Armada's value is not automatic retry — the
NRP docs state plainly that
"preempted jobs will not be automatically rescheduled", and armadactl get retry-policies
returns Unimplemented on NRP. Reaching for Armada to survive preemption is the wrong reason.
Because nothing retries, gap-fill is a pipeline stage rather than an exception — measured 1-4 missing slices per 3,780 (~0.1%). See "Closing the retry loop" below; you no longer have to write the gap-fill by hand for a raster hex build.
The real reason is that Armada is not bound by the k8s indexed-Job completion cap (~200, an etcd pressure limit). Millions of completions are fine. That makes microslicing practical, and microslicing improves four things at once:
| few large jobs | many small slices | |
|---|---|---|
| unit of loss on preemption | hours | minutes |
| RAM requested | sized for the worst step in the chain | what each step actually needs |
| scheduling | needs large contiguous free slots; pods sit Pending |
fits in scraps almost anywhere |
| stragglers | one holds up the batch and can leave stale output | retries invisibly |
The scheduling point is the one most easily missed: 20 pods at 64Gi need ~1.3 TB free in large contiguous chunks, so realized parallelism is far below requested. Thousands of 8-16Gi slices pack into whatever is free and start immediately. Placing 10,000 small slices is often faster than placing 100 big ones — more so under preemption.
Big jobs also make you conservative: a 35-step chain must request the peak requirement of its worst step, so every step pays that cost. One unit per job asks for what it needs.
Microslicing a raster hex is now generated, not hand-written
cng-datasets raster-workflow emits sub-h0 units directly, so the CHELSA-era pattern of a
bespoke gen_armada_*.py per dataset is no longer the only route for the spatial dimension:
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.
- 10d ago Changed · +7 lines 6363f20cbef8
- 1mo ago Changed · +129 lines · +23 tokens per session 8c410a6d2193
- 2mo ago First seen · 292 lines · 107 tokens per session scan A 628b90075601
armada-pipeline is a skill published in the GitHub repository boettiger-lab/data-workflows (5 stars, last pushed 2d ago), licensed BSD-3-Clause. It adds 130 tokens to every session and 5,644 once invoked, about $0.0005 per session on Opus 5.5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
armada-pipeline compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| armada-pipelineboettiger-lab/data-workflows | Skill | 5 | 2d ago | BSD-3-Clause | A | 130 | Claude Code |
| nextflowK-Dense-AI/scientific-agent-skills | Skill | 56k | 4d ago | MIT | A | 74 | Claude Code, Codex |
| latchbio-integrationK-Dense-AI/scientific-agent-skills | Skill | 56k | 4d ago | MIT | A | 78 | Claude Code |
| remote-compute-sshaipoch/open-science | Skill | 5.5k | today | Apache-2.0 | A | 53 | Claude Code, Codex |
| modal-computecompanion-inc/feynman | Skill | 10k | 6d ago | MIT | A | 39 | Claude Code, Codex |
| dnanexus-integrationsynthetic-sciences/openscience | Skill | 4.3k | 2d ago | Apache-2.0 | A | 49 | Claude Code, Codex |
| latchbio-integrationsynthetic-sciences/openscience | Skill | 4.3k | 2d ago | Apache-2.0 | A | 45 | Claude Code, Codex |