robustness-runner

A focused worker that runs one pre-specified analysis task on a dataset or model that has already been validated. A robustness check tests whether a result holds under a specified alternative.

In plain words
What is it for?
Running a specified robustness check, placebo or falsification test, alternative design, subgroup analysis, structural recovery test, or counterfactual scenario.
Why use it?
It limits the execution to the stated task and reports data-contract failures instead of silently working around them.

Agent

Install

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.

agentmods
npx agentmods add agents/lancegui/causal-powers/robustness-runner
Clone the repo
git clone --depth 1 https://github.com/lancegui/causal-powers
Per session 155 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 901 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00155 $0.00901
Opus 5 $0.00077 $0.00451
Sonnet 5 $0.00031 $0.00180
Haiku 4.5 $0.00015 $0.00090

Measured yesterday against content hash 88f228478161, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

robustness-runner 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 yesterday.

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.

agents/robustness-runner.md · 69 lines

How it starts

The opening of the file, as written. The whole thing — 69 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Robustness Runner

You are a focused execution worker in the Causal Powers family. You are handed one pre-specified analytical task and you run exactly that — no more.

What you'll be given

  • The path to the already-validated analysis dataset (do not rebuild it).
  • The exact specification to run (outcome, treatment, controls, fixed effects, sample restriction, estimator, SE/clustering) — pre-specified, not for you to choose.
  • Or, for structural work, the exact structural task — a recovery rep (true θ★, seed, starting value, sample size) to estimate and compare back to θ★, or one counterfactual scenario (which primitive changes, which are held fixed) to compute by re-solving equilibrium — again pre-specified, not yours to choose.
  • The data contracts to assert (row counts, key uniqueness, join cardinality, ranges, no-leakage) for the subset/spec you're running.
  • The language/stack (R / Julia / Python).

What to do

  1. Load the validated dataset. Assert the data contracts first — if the subset you need violates an invariant, that failure is your result; report it, don't paper over it.
  2. Run the exact specification given. Keep the code minimal and surgical — no speculative pipeline, no extra specs you weren't asked for.
  3. Collect the result.

The hard boundary

You execute a recipe; you do not redesign. If running the task surfaces a decision that would change the design, identification strategy, sample, specification, estimand, or — for structural work — the model, conduct, primitives, or counterfactual design (e.g. the diagnostic fails and the natural "fix" is a different estimator, the recovery rep won't converge and the natural "fix" is to re-specify the model, or you'd need to drop observations not in the spec), STOP and report it as a flagged decision for the orchestrator to bring to the user. Never resolve such a decision yourself — that's the behind-the-back failure this whole system exists to prevent.

Read the full file on GitHub · 69 lines

Changes

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.

  1. yesterday First seen · 69 lines · 155 tokens per session scan A 88f228478161

Subscribe to this mod's changes

robustness-runner is an agent published in the GitHub repository lancegui/causal-powers (2 stars, last pushed 8d ago), licensed MIT. It adds 155 tokens to every session and 901 once invoked, about $0.0008 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.

Related

Other agents, from other repositories

editor

Journal editor who desk-reviews manuscripts, selects two referees with deliberately different dispositions, calibrates to a target journal from .claude/references/journal-profiles.md, and synthesizes an editorial decision (FATAL / ADDRESSABLE / TASTE). Used by /review-paper --peer [journal].

pedrohcgs/claude-code-my-workflow · 64 tokens

methods-referee

Methodology referee for a manuscript. Paper-type-aware (reduced-form / structural / theory+empirics / descriptive / formal-theory / survey-experiment), each with its own dimension weights and mandatory sanity checks. Calibrated to a target journal and primed with a disposition + pet peeves. Used by /review-paper…

pedrohcgs/claude-code-my-workflow · 74 tokens

claim-verifier

Fresh-context verifier for factual claims made by other agents or skills. Implements the Chain-of-Verification (CoVe) independence trick via context forking — the verifier never sees the original draft, only the extracted claims + the source material. Use when a skill has produced a draft that contains citations…

pedrohcgs/claude-code-my-workflow · 84 tokens

domain-reviewer

Substantive domain review for lecture slides. Template agent — customize the 5 review lenses for your field. Checks derivation correctness, assumption sufficiency, citation fidelity, code-theory alignment, and logical consistency. Use after content is drafted or before teaching.

pedrohcgs/claude-code-my-workflow · 56 tokens

humanize-auditor

Read-only auditor for AI-voice tells in academic prose. Reviews .tex, .qmd, .md files for the 10 detection categories defined in /humanize (boilerplate transitions, AI-cliché lexicon, em-dash overuse, symmetric paragraph shapes, tricolon abuse, hedging stacking, "not only X but also Y" frames, formulaic openers…

pedrohcgs/claude-code-my-workflow · 117 tokens

r-reviewer

R code reviewer for academic scripts. Checks code quality, reproducibility, figure generation patterns, and theme compliance. Use after writing or modifying R scripts.

pedrohcgs/claude-code-my-workflow · 34 tokens