Skip to content
Back to search

high-stakes-analytics-decision-lab

A platform-neutral analytical Skill that profiles messy data, selects case-adaptive methods, and produces source-backed visual reports for high-stakes decisions.

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today
Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/limingrui679-design--high-stakes-analytics-decision-lab/badge.svg)](/listings/limingrui679-design--high-stakes-analytics-decision-lab)

HTML

<a href="/listings/limingrui679-design--high-stakes-analytics-decision-lab"><img src="/api/listings/limingrui679-design--high-stakes-analytics-decision-lab/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

npx skills add limingrui679-design/high-stakes-analytics-decision-lab -g

The Agent Skills installer discovers the compact package under [`skills/high-stakes-analytics-decision-lab/`](skills/high-stakes-analytics-decision-lab/): 39 files and about 472 KiB, rather than the full research portfolio. Its machine-readable file and hash contract is in [`bundle-manifest.json`](skills/high-stakes-analytics-decision-lab/bundle-manifest.json).

Use [`docs/getting-started.md`](docs/getting-started.md) for Codex-specific, no-install, and direct repository options.

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

README.md · README · 275 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. <p align="center">
  2. <img src="assets/readme-hero.svg" alt="High-Stakes Analytics & Decision Lab routes a question to an Evidence Intelligence Report and adds a Decision Intelligence Brief only when justified" width="100%">
  3. </p>
  4. <h1 align="center">High-Stakes Analytics & Decision Lab</h1>
  5. <p align="center">
  6. Evidence-gated analytics for consequential questions.<br>
  7. Move from an ambiguous decision to reproducible evidence—and only then, when justified, to bounded action.
  8. </p>
  9. <p align="center">
  10. <a href="https://github.com/limingrui679-design/high-stakes-analytics-decision-lab/releases/tag/v1.1.2"><img src="https://img.shields.io/badge/release-v1.1.2-008C82?style=flat-square" alt="Release v1.1.2"></a>
  11. <a href="https://github.com/limingrui679-design/high-stakes-analytics-decision-lab/actions/workflows/verify.yml"><img src="https://github.com/limingrui679-design/high-stakes-analytics-decision-lab/actions/workflows/verify.yml/badge.svg" alt="Verification status"></a>
  12. <a href="https://github.com/limingrui679-design/high-stakes-analytics-decision-lab/actions/workflows/codeql.yml"><img src="https://github.com/limingrui679-design/high-stakes-analytics-decision-lab/actions/workflows/codeql.yml/badge.svg" alt="CodeQL status"></a>
  13. <a href="LICENSE.txt"><img src="https://img.shields.io/badge/license-MIT-6B7280?style=flat-square" alt="MIT License"></a>
  14. </p>
  15. <p align="center">
  16. <a href="#start-in-three-steps"><strong>Quick start</strong></a> ·
  17. <a href="#how-it-works"><strong>Architecture</strong></a> ·
  18. <a href="https://limingrui679-design.github.io/high-stakes-analytics-decision-lab/demo/"><strong>Live explorer</strong></a> ·
  19. <a href="#fifteen-complete-evidence-paths"><strong>Portfolio</strong></a> ·
  20. <a href="docs/README.md"><strong>Documentation</strong></a> ·
  21. <a href="CONTRIBUTING.md"><strong>Contributing</strong></a>
  22. </p>
  23. ## Why this exists
  24. High-stakes analysis often fails before the model: the question is underspecified,
  25. the data contract is implicit, cleaning choices are hidden, uncertainty is treated
  26. as independent, or a recommendation is written because the template expects one.
  27. This repository is a platform-neutral Agent Skill and reproducible research
  28. portfolio built around a stricter sequence:
  29. | Principle | System behavior |
  30. |---|---|
  31. | **Evidence before method** | Declare the question, population, grain, target quantity, horizon, lineage, and claim boundary first |
  32. | **Readiness before analysis** | Preserve the source, profile quality and privacy, and pause on material transformations |
  33. | **Adaptive routes** | Add descriptive, diagnostic, predictive, or prescriptive work only when justified |
  34. | **Honest endpoints** | Accept an evidence request, negative validation, `do_not_deploy`, or no recommendation |
  35. | **Dependent uncertainty** | Retain shared time, market, participant, campaign, operational, and spatial shocks |
  36. | **Traceable communication** | Link claims and accessible figures to JSON, CSV, hashes, and rerunnable code |
  37. The result is not a fixed report generator. It is an evidence-gated orchestration
  38. system that can stop, ask for a named decision, or produce a bounded analytical
  39. product without upgrading weak evidence into a stronger claim.
  40. ## Start in three steps
  41. ### 1. Install the Skill
  42. ```bash
  43. npx skills add limingrui679-design/high-stakes-analytics-decision-lab -g
  44. ```
  45. The Agent Skills installer discovers the compact package under
  46. [`skills/high-stakes-analytics-decision-lab/`](skills/high-stakes-analytics-decision-lab/):
  47. 39 files and about 472 KiB, rather than the full research portfolio. Its
  48. machine-readable file and hash contract is in
  49. [`bundle-manifest.json`](skills/high-stakes-analytics-decision-lab/bundle-manifest.json).
  50. Use [`docs/getting-started.md`](docs/getting-started.md) for Codex-specific,
  51. no-install, and direct repository options.
  52. ### 2. Ask for the evidence outcome
  53. ```text
  54. $high-stakes-analytics-decision-lab
  55. Run the data-readiness gate on this source, preserve the original file, and
  56. select only the analytical routes the evidence supports. Produce an Evidence
  57. Intelligence Report. Add a Decision Intelligence Brief only if the evidence
  58. and decision context justify one.
  59. ```
  60. Start with the decision or evidence question—not a preferred model. A valid
  61. result may be a bounded action, a pilot requirement, targeted diligence, an
  62. evidence request, negative validation, or `do_not_deploy`.
  63. ### 3. Review the evidence package
  64. Every complete project keeps the narrative, machine result, visual evidence,
  65. and source lineage together:
  66. ```text
  67. report.md # primary Evidence Intelligence Report
  68. results.json # machine-readable analytical result
  69. chart-map.json # figure-to-question and source contract
  70. figures/*.svg # accessible analytical visuals
  71. ```
  72. A justified decision layer adds `decision-report.md`,
  73. `decision-results.json`, and its own figure contract. It never replaces the
  74. primary evidence product.
  75. ### Direct repository entry points
  76. | Starting point | Command or guide | Outcome |
  77. |---|---|---|
  78. | Environment audit | `python3 scripts/hsadl.py doctor` | Python, runtime, template, write-access, and Skill-footprint checks |
  79. | Safe 60-second walkthrough | `python3 scripts/hsadl.py demo --output-dir build/demo` | Synthetic source preservation, contract, quality gate, route, and accessible SVGs; no model or recommendation |
  80. | Question only | `python3 scripts/hsadl.py route "<question>" --scope full --output-dir <path>` | Evidence and method blueprint; no invented result |
  81. | Question plus data | `python3 scripts/hsadl.py start <data.csv> --question "<question>" --output-dir <path>` | Preserved source, draft contract, readiness profile, and unresolved decisions |
  82. | Existing decision case | `python3 scripts/hsadl.py validate <case.json>` then `python3 scripts/hsadl.py run <case.json> --output-dir <path>` | Validated expected, tail, sensitivity, provenance, and group-impact outputs |
  83. | Worked precedents | [Fifteen-project portfolio](examples/real-data-cases/README.md) | Complete source-to-report evidence paths |
  84. ## How it works
  85. <p align="center">
  86. <img src="assets/adaptive-reporting-system.svg" alt="Adaptive reporting routes the case before choosing fields, methods, figures, and terminal status" width="82%">
  87. </p>
  88. The fixed evidence spine remains stable while the case-specific analytical
  89. layer changes.
  90. | Fixed evidence spine | Adaptive case layer |
  91. |---|---|
  92. | Question, population, unit, target quantity, and horizon | Route, fields, methods, and validation |
  93. | Source lineage, quality status, and reproducibility | Figures, report sections, and decision criteria |
  94. | Uncertainty, limitations, and claim boundary | Bounded action, evidence request, or stopping status |
  95. ### The data gate can stop the workflow
  96. Uploaded row-level data do not go directly into a model. The system preserves
  97. the original, establishes a contract, checks grain and keys, profiles quality
  98. and privacy, and produces a dry-run remediation plan.
  99. | Gate status | Meaning | Permitted next step |
  100. |---|---|---|
  101. | `ready` | No material failure under the declared contract | Continue |
  102. | `ready_with_documented_limitations` | Localized issues remain | Continue with visible limits |
  103. | `needs_user_confirmation` | A substantive transformation, privacy, or intended-use choice remains | Pause for a named approval or clarification |
  104. | `blocked` | Grain, key, schema, leakage, or another critical failure invalidates the route | Stop and request corrected evidence |
  105. Only safe normalization can run without approval. Deletion, imputation,
  106. outlier treatment, category merging, unit conversion, target correction, and
  107. grain changes require explicit action IDs. The processed copy never overwrites
  108. the source.
  109. ### Four routes, no mandatory recommendation
  110. | Route | Question | Required discipline | Valid endpoint |
  111. |---|---|---|---|
  112. | **Descriptive** | What is happening? | Denominators, coverage, trends, segments, and missingness | Baseline report or evidence request |
  113. | **Diagnostic** | Why might it be happening? | Contributions, competing explanations, hypotheses, and a visible causal boundary | Prioritized explanations to test |
  114. | **Predictive** | What is likely next? | Target, horizon, baseline, held-out validation, calibration, subgroup error, and drift | Validated prediction, negative validation, or `do_not_deploy` |
  115. | **Prescriptive** | What should be done, if justified? | Owner, alternatives, constraints, dependence, tail risk, sensitivity, and reversal conditions | Bounded action or no decision-ready recommendation |
  116. Routes may compose, but a later route cannot erase the quality and evidence
  117. requirements of an earlier stage. Read the full system design in
  118. [`docs/architecture.md`](docs/architecture.md).
  119. ## Two products, one evidence contract
  120. ![The Evidence Intelligence Report is primary and the Decision Intelligence Brief is conditional](assets/report-layers.svg)
  121. | Product | Main question | Contents | Existence rule |
  122. |---|---|---|---|
  123. | **Evidence Intelligence Report** | What does the evidence establish? | Source and QA contract, methods, validation, figures, uncertainty, limitations, lineage, and reproducibility | Primary record for every complete project |
  124. | **Decision Intelligence Brief** | What action, pilot, diligence, evidence request, or stop follows? | Decision status, alternatives, constraints, shared shocks, tail risk, sensitivity, and reversal conditions | Conditional; only when a separate decision layer is justified |
  125. Every material figure is generated from the shared editorial evidence system,
  126. includes a title and description, and is paired with its analytical question,
  127. supported interpretation, and claim boundary. The visual system is part of the
  128. evidence contract, not decorative reporting.
  129. ## Fifteen complete evidence paths
  130. <p align="center">
  131. <img src="examples/real-data-cases/figures/case-landscape.svg" alt="Fifteen real-data cases and their evidence-matched analytical paths" width="92%">
  132. </p>
  133. The public portfolio contains 15 primary reports and 10 conditional briefs—25 intelligence products in total—plus 119 canonical accessible figures: 50 evidence figures and 69 decision figures.
  134. The cases span operational demand, distribution shift, scarce-capacity pilots,
  135. temporal model transport, tail-risk decisions, human-in-the-loop triage,
  136. commercial diligence, mitigation allocation, filing review, clustered field
  137. experiments, survival evidence, policy evaluation, repeated-measures inference,
  138. population transportability, and spatial equity.
  139. They intentionally end differently: some support a bounded decision, some
  140. require a pilot or targeted review, and some stop at an evidence request or
  141. `do_not_deploy`.
  142. | Explore | Open |
  143. |---|---|
  144. | Searchable route and capability explorer | [Live case atlas](https://limingrui679-design.github.io/high-stakes-analytics-decision-lab/demo/) · [local source](demo/index.html) |
  145. | Visual case gallery | [Portfolio overview](examples/real-data-cases/README.md) |
  146. | Machine-readable catalog | [`cases.json`](examples/real-data-cases/cases.json) |
  147. | School-neutral capability paths | [`capability-map.json`](examples/real-data-cases/capability-map.json) |
  148. | Rebuild and comparison contract | [Verification guide](docs/verification.md) |
  149. | Method and domain routing | [`method-domain-map.json`](references/method-domain-map.json) |
  150. ## Verification you can reproduce
  151. The current stable release is
  152. [`v1.1.2`](https://github.com/limingrui679-design/high-stakes-analytics-decision-lab/releases/tag/v1.1.2).
  153. Its versioned source package and SHA-256 checksum are published together; the
  154. release identity is also recorded in [`CITATION.cff`](CITATION.cff),
  155. [`CHANGELOG.md`](CHANGELOG.md), and `RELEASE-MANIFEST.json`.
  156. ```bash
  157. make verify
  158. ```
  159. This runs the standalone regression suite and rebuilds all fifteen projects in
  160. an isolated verified-file copy. Use `make quality` for the tracked-secret scan,
  161. Ruff, mypy, and codespell; use [`docs/verification.md`](docs/verification.md)
  162. for direct commands, static security gates, and no-Git release verification.
  163. The 106 public tests cover data readiness, adaptive routing, numerical behavior,
  164. source and artifact identity, package integrity, local links, accessible SVGs,
  165. generator idempotence, compact installation, quickstart safety, interactive
  166. explorer synchronization, no-Git releases, DNS and SSRF boundaries, and source
  167. parser security. The complete 106-test standalone regression suite and the
  168. fifteen-project rebuild are separate gates: successful reproduction establishes
  169. the reviewed workflow and declared numerical tolerance, not empirical validity,
  170. external adoption, or real-world impact.
  171. CI exercises Python 3.11, 3.12, 3.13, and 3.14. Security checks include the
  172. tracked-secret gate, Bandit, dependency auditing, hardened source-builder tests,
  173. targeted branch coverage, and CodeQL.
  174. ## Repository design
  175. ```text
  176. high-stakes-analytics-decision-lab/
  177. ├── skills/high-stakes-analytics-decision-lab/
  178. │ └── SKILL.md # compact, installable Agent Skill package
  179. ├── demo/ # dependency-free interactive case explorer
  180. ├── docs/ # onboarding, architecture, layout, verification
  181. ├── references/ # enforceable method and evidence contracts
  182. ├── assets/ # templates and canonical README visuals
  183. ├── scripts/ # profiling, routing, analysis, generation, checks
  184. ├── examples/real-data-cases/ # fifteen reproducible evidence projects
  185. ├── tests/ # standalone contract and regression suite
  186. └── .github/ # CI, security, issue, and PR workflows
  187. ```
  188. The nested Skill package prevents compatible installers from copying the full
  189. portfolio. It is generated from the canonical root scripts, references, and
  190. templates by `scripts/build_skill_bundle.py`; the full cases remain in
  191. `examples/`. Human guides live in `docs/`; precise runtime rules live in
  192. `references/`; generated evidence stays beside the project that produced it.
  193. See [`docs/repository-layout.md`](docs/repository-layout.md) before moving or
  194. regenerating files.
  195. ## Documentation
  196. | Guide | Use it for |
  197. |---|---|
  198. | [Documentation home](docs/README.md) | Choose a user, reviewer, or maintainer path |
  199. | [Getting started](docs/getting-started.md) | Install, route a question, profile data, and run a case |
  200. | [Architecture](docs/architecture.md) | Understand gates, routes, products, the decision engine, and visual evidence |
  201. | [Repository layout](docs/repository-layout.md) | Distinguish runtime contracts, human guides, canonical sources, and generated files |
  202. | [Verification](docs/verification.md) | Reproduce tests, portfolio outputs, quality checks, and release gates |
  203. | [Contributing](CONTRIBUTING.md) | Change code, documentation, sources, or cases without breaking evidence boundaries |
  204. | [Security](SECURITY.md) | Report a vulnerability privately |
  205. ## Responsible-use boundary
  206. This is a public, tested research and portfolio prototype. It does not establish
  207. production readiness, institutional adoption, medical advice, investment
  208. advice, a regulatory finding, an assurance opinion, or achieved real-world
  209. impact. A reproducible result can still be decision-inappropriate; domain
  210. review remains mandatory before operational use.
  211. ## Citation and license
  212. Cite the version reviewed using [`CITATION.cff`](CITATION.cff). Public release
  213. history is in [`CHANGELOG.md`](CHANGELOG.md), and component-version boundaries
  214. are in [`VERSIONING.md`](VERSIONING.md).
  215. Licensed under the [MIT License](LICENSE.txt).

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.