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 agentmods add skills/lancegui/causal-powers/using-causal-powersnpx skills add lancegui/causal-powers --skill using-causal-powersgit clone --depth 1 https://github.com/lancegui/causal-powersWhat 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 | $0.00202 | $0.04223 |
| Opus 5 | $0.00101 | $0.02112 |
| Sonnet 5 | $0.00040 | $0.00845 |
| Haiku 4.5 | $0.00020 | $0.00422 |
Grade A, and why
using-causal-powers 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 2d 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 — 123 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Using Causal Powers
The creed
A number you computed but never validated is a guess wearing a lab coat. In software the dangerous bug throws a stack trace; in data analysis it runs clean and hands you a confident, wrong answer. Causal Powers is a family of skills that make the silent failures of data work loud — before they reach a stakeholder.
The rule
For any analysis task, invoke the relevant discipline skill before acting — before exploring the data, before writing the transform, before reporting the number. Process skills (framing the question, planning, debugging) come before implementation. Even a 1% chance a skill applies means you invoke the Skill tool to check (read it, then decide) — not just "consider it in your head."
Re-trigger per request — a skill you used earlier does not stay satisfied. Every new ask re-fires the relevant skill, even on an already-locked, already-reviewed design. A re-run or a finer reporting cut is still executing-analysis-plans (run the approved plan, fan independent work to subagents), with result-verification offered before you write any result to a file — it runs only if the user asks; "review it" re-fires analysis-review; a cut that changes the unit or estimand re-opens question-framing + analysis-checkpoints. "I already have the context" / "this is just running the locked plan" is the rationalization that skips the gate — and it's exactly how a reviewed design ships an unverified new cut. But re-firing means re-applying the discipline, not reflexively reloading the file: if the skill is still in this session's context, apply it and say so; re-invoke the Skill tool only when its body has scrolled out, was compacted away, or you need the details — the discipline is mandatory, the reload is not (it's wasted context to reload a skill you already hold).
And the rule the rest of the family rests on: you execute autonomously toward the agreed goal, but never change it behind the user's back — the design, estimand, sample, spec, or a metric. When execution wants to (it most often does mid-debugging), STOP and bring it to them (analysis-checkpoints owns the full checkpoint list and the write→sign-off→reconstruct mechanics behind every locked plan). This is the discipline missing when an analysis quietly became one nobody agreed to.
What ships with it
3 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 2d ago First seen · 123 lines · 202 tokens per session scan A 0bacd1b188d3
using-causal-powers is a skill published in the GitHub repository lancegui/causal-powers (2 stars, last pushed 8d ago), licensed MIT. It adds 202 tokens to every session and 4,223 once invoked, about $0.0010 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 2026-08-31.
Other skills, from other repositories
audit-reproducibility
Enforce the replication-protocol.md rule by cross-checking numeric claims in a manuscript against the actual R / Stata / Python outputs. Report PASS/FAIL per claim against tolerance thresholds. Use before submission and before releasing a replication package.
diagnose
Root-cause a failing or wrong empirical result with a disciplined reproduce → minimise → hypothesise → instrument → fix loop, instead of guessing-and-poking. Use when the user says "why is my regression wrong", "this number changed", "my script errors out", "the result won't reproduce", "debug this", "this estimate…
capture-environment
Snapshot the computational environment for a replication package — detects the analysis stack (R / Stata / Python) and emits the right lockfiles (renv.lock + sessionInfo.txt, requirements.txt / environment.yml / uv.lock, Stata version + ado package list), records seeds and RNG kind, optionally writes a pinning…
checkpoint
Save a structured state snapshot before stopping or handing off. Captures the active plan, recent decisions, file pointers (with line numbers), open questions, and the next 1–3 actions into a checkpoint file under qualityreports/checkpoints/. Optionally proposes [LEARN] entries to add to MEMORY.md. Use when user says…
coauthor-brief
Generate a co-author / collaborator handoff brief for a multi-author, multi-machine project — summarizing what changed since the last brief (git delta), the current state of each artifact (manuscript, analysis, slides), open questions, how to reproduce locally, and any restricted-data access steps. Use when user says…
data-management-plan
Draft a funder-compliant Data Management Plan (NSF DMP, NIH DMS Policy 2023, ERC, Horizon Europe) by composing the confidential-data and environment-capture primitives. Sections cover data description, formats/metadata, storage/backup, access/sharing, preservation/archiving, and roles. Use when user says "data…