Skip to content
Back to search

academic-research-skills-codex

Codex-native Academic Research Skills suite for human-in-the-loop academic research workflows

Safety scan

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

Checked today

Check it deeper on your machine

Its scripts were only partly scanned.

As a Claude Code plugin, install it and then ask Claude Code:

/plugin marketplace add keithmackay/mackayi/plugin install scanskillsafety@mackayiDo a deep scan of https://github.com/imbad0202/academic-research-skills-codex with scanskillsafety

Or from a terminal, clone it and run it:

git clone https://github.com/keithmackay/scanskillsafety-skill ~/.claude/skills/scanskillsafetynode ~/.claude/skills/scanskillsafety/cli.cjs --deep 'https://github.com/imbad0202/academic-research-skills-codex'

It checks the exact commit: every file, its packages against OSV (api.osv.dev), how it compares with this page's scan (findsafeskills.com receives the repository's address), and, if you agree, Claude reading every file with your own Claude subscription. It never runs the code.

Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/imbad0202-academic-research-skills-codex/badge.svg)](/listings/imbad0202-academic-research-skills-codex)

HTML

<a href="/listings/imbad0202-academic-research-skills-codex"><img src="/api/listings/imbad0202-academic-research-skills-codex/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.

Add the GitHub marketplace and install ARS-Codex with Codex CLI:

codex plugin marketplace add Imbad0202/academic-research-skills-codex --ref main
codex plugin add ars-codex@ars-codex

To update a plugin install later:

codex plugin marketplace upgrade ars-codex
codex plugin add ars-codex@ars-codex

In Codex Desktop, you can alternatively add the repository from **Plugins** and then install **ARS-Codex**:

Marketplace source: https://github.com/Imbad0202/academic-research-skills-codex.git
Branch/ref: main
Plugin: ars-codex

The plugin root is `plugins/ars-codex/`. Its `skills/` directory contains a materialized copy of `academic-research-suite`, not a symlink. This keeps Codex Desktop installs portable on Windows, where plugin caches may materialize symlinks as plain text files and skip bundled skill registration.

Open a new Codex conversation after installation, then invoke `$academic-research-suite` or describe an academic research task that matches the bundled workflow.

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 · 579 lines

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

This file has changed since it was scanned, so this is not exactly the text that was scanned, and no line is marked. The findings are listed with the safety scan above.

  1. # ARS-Codex
  2. [![Version](https://img.shields.io/badge/version-v3.24.2-blue)](VERSION)
  3. [![License: CC BY-NC 4.0](https://img.shields.io/badge/license-CC%20BY--NC%204.0-lightgrey)](https://creativecommons.org/licenses/by-nc/4.0/)
  4. [![DOI](https://img.shields.io/badge/DOI-10.5281%2Fzenodo.20784820-blue)](https://doi.org/10.5281/zenodo.20784820)
  5. [![Sponsor](https://img.shields.io/badge/sponsor-Buy%20Me%20a%20Coffee-orange?logo=buy-me-a-coffee)](https://buymeacoffee.com/crucify020v)
  6. **Languages:** [English](README.md) · [简体中文](README_ZH-CN.md) · [繁體中文](README_ZH-TW.md) · [日本語](README_JA.md)
  7. ARS-Codex is the Codex-native sibling of
  8. [Academic Research Skills (ARS) for Claude Code](https://github.com/Imbad0202/academic-research-skills).
  9. It is a separate Codex distribution with its own plugin identity, packaging,
  10. and runtime adapter. Its release number follows the vendored ARS suite.
  11. This repository vendors the ARS workflow content as a single Codex skill:
  12. ```text
  13. skills/academic-research-suite/
  14. SKILL.md
  15. manifest.json
  16. agents/openai.yaml
  17. codex/
  18. full-runtime-manifest.json
  19. agents/
  20. hooks/
  21. scripts/
  22. ars/
  23. deep-research/
  24. academic-paper/
  25. academic-paper-reviewer/
  26. academic-pipeline/
  27. sr-screener/
  28. experiment-agent/
  29. commands/
  30. hooks/
  31. docs/
  32. tests/
  33. shared/
  34. ```
  35. The original Claude Code ARS checkout is not modified. Upstream content is copied
  36. from verified upstream release commits and adapted through the Codex router in
  37. `skills/academic-research-suite/SKILL.md`.
  38. ## Relationship to Claude Code ARS
  39. This repository is ARS-Codex. For the original Claude Code ARS distribution, use
  40. [Imbad0202/academic-research-skills](https://github.com/Imbad0202/academic-research-skills).
  41. Use the Claude Code repo when you want the native Claude Code skill layout,
  42. Claude-specific agent-team behavior, or the original ARS development history.
  43. Use this repo when you want the Codex-native single-suite skill.
  44. ## Versioning
  45. This ARS-Codex package is version `3.24.2`. The repo-root `VERSION` file,
  46. `skills/academic-research-suite/SKILL.md` metadata version, and
  47. `skills/academic-research-suite/manifest.json` `adapter_version` track the
  48. Codex package version in step with the vendored ARS suite, starting at `3.22.0`.
  49. Earlier `0.1.x` release numbers remain historical. The exact upstream version,
  50. tag, and commit are recorded in `manifest.source_repositories[]`.
  51. Package-level changes are summarized in [`CHANGELOG.md`](CHANGELOG.md).
  52. The vendored ARS source tracks **v3.24.2** at
  53. `Imbad0202/academic-research-skills@fd5ff1b858374219b4d342e38587b66e18f61cbd`.
  54. This release carries upstream v3.24.1 and v3.24.2. The reviewer now stops after
  55. Phase 0: it shows every Reviewer Configuration Card and starts no reviewer seat
  56. until you accept the set in your own turn, with or without edits. Only your own
  57. request to skip the card confirmation skips it. The rule is prompt-level, and
  58. whether Codex sessions stop at it is unmeasured. An experiment with no honest lock
  59. values can use `repro_lock: null` with a reason; earlier releases reject that form.
  60. `check_repro_lock.py` is stricter, so some passports that passed before now fail.
  61. The sr-screener methods sentence is corrected; text generated by an earlier release
  62. keeps the old sentence, so correct it by hand. The upstream claude.ai install,
  63. plugin agent-name, and legacy-name guard changes remain inactive Claude references
  64. in Codex. There is no new workflow, and the native Codex model policy is unchanged.
  65. The output-language-pair registry still supports only `zh-tw-en`; Spanish
  66. triggers do not install a Spanish output-locale pack. Claude model audits and
  67. plugin eval suites are reference material and do not change the Codex model
  68. policy or establish measured Codex performance.
  69. Nested upstream `.github/` workflows and root `agents/` mirrors are preserved
  70. for traceability and self-tests, but are not repo-level CI or Codex entrypoints;
  71. Claude/plugin loader files under `.claude/` and `.claude-plugin/` remain
  72. intentionally excluded.
  73. ## Models and execution
  74. This checkout's [project configuration](.codex/config.toml) selects `gpt-6-astra`
  75. with `xhigh` reasoning for new trusted Codex sessions. Installing the skill does
  76. not copy that project setting or change an already-running session. The planner
  77. suggests `medium` for routine work and `xhigh` for complex work; these are
  78. starting policies, not measured optimal settings. `ultra` is optional in Codex
  79. for explicitly chosen demanding work and is rejected by the contained citation
  80. transport.
  81. Native adaptive execution is the default: use bounded subagents for independent
  82. work when useful and let the lead continue with other work. The fixed
  83. full-runtime topology and hooks remain opt-in. Read the
  84. [model runtime policy](skills/academic-research-suite/codex/model-runtime-policy.md)
  85. and [system-card alignment audit](skills/academic-research-suite/codex/audits/2026-09-06-model-alignment.md).
  86. ## Install ARS-Codex Plugin
  87. Add the GitHub marketplace and install ARS-Codex with Codex CLI:
  88. ```bash
  89. codex plugin marketplace add Imbad0202/academic-research-skills-codex --ref main
  90. codex plugin add ars-codex@ars-codex
  91. ```
  92. To update a plugin install later:
  93. ```bash
  94. codex plugin marketplace upgrade ars-codex
  95. codex plugin add ars-codex@ars-codex
  96. ```
  97. **Windows update recovery:** if failed updates repeatedly create
  98. `marketplace-upgrade-*` directories or fill the system drive, stop retrying and
  99. open the [local marketplace recovery guide](docs/windows-marketplace-recovery.md).
  100. If the drive is full, first follow its step 1 to stop Codex and free space
  101. safely. Codex CLI 0.164.0 includes upstream fixes that clean up leftover staging
  102. directories, so then check the CLI version bundled with Codex Desktop and update
  103. if it is older. If leftover directories still accumulate on 0.164.0 or later,
  104. continue with the guide: it covers preserving a working copy, switching only ARS
  105. to a local source, verification, and manual updates. After switching to a local source, use that
  106. guide's manual update procedure instead of `marketplace upgrade`. Disabling the
  107. plugin alone is insufficient: Codex can still refresh files for disabled
  108. configured plugins. The guide is a workaround for [#43](https://github.com/Imbad0202/academic-research-skills-codex/issues/43);
  109. the Windows `os error 5` failure reported there has not been retested on 0.164.0.
  110. In Codex Desktop, you can alternatively add the repository from **Plugins** and
  111. then install **ARS-Codex**:
  112. ```text
  113. Marketplace source: https://github.com/Imbad0202/academic-research-skills-codex.git
  114. Branch/ref: main
  115. Plugin: ars-codex
  116. ```
  117. The plugin root is `plugins/ars-codex/`. Its `skills/` directory contains a
  118. materialized copy of `academic-research-suite`, not a symlink. This keeps
  119. Codex Desktop installs portable on Windows, where plugin caches may materialize
  120. symlinks as plain text files and skip bundled skill registration.
  121. Open a new Codex conversation after installation, then invoke
  122. `$academic-research-suite` or describe an academic research task that matches
  123. the bundled workflow.
  124. ## Direct Skill Install Or Update
  125. As an alternative to the plugin, install the skill directly from this repo
  126. path. Use `--method git` so public and
  127. credentialed GitHub access both work consistently:
  128. ```bash
  129. python3 "$HOME/.codex/skills/.system/skill-installer/scripts/install-skill-from-github.py" \
  130. --repo Imbad0202/academic-research-skills-codex \
  131. --ref main \
  132. --path skills/academic-research-suite \
  133. --method git
  134. ```
  135. On macOS and many Linux systems, Python 3 is exposed as `python3` rather than
  136. `python`. If your system only has a `python` command and it is Python 3, use
  137. `python` in the commands instead.
  138. To update an existing install:
  139. ```bash
  140. rm -rf "$HOME/.codex/skills/academic-research-suite"
  141. python3 "$HOME/.codex/skills/.system/skill-installer/scripts/install-skill-from-github.py" \
  142. --repo Imbad0202/academic-research-skills-codex \
  143. --ref main \
  144. --path skills/academic-research-suite \
  145. --method git
  146. ```
  147. Open a new Codex conversation after installation. Existing Codex sessions may
  148. keep their old skill cache; you do not need to close unrelated Claude or Codex
  149. sessions.
  150. Verify with `/skills`: you should see one ARS-Codex entry,
  151. `academic-research-suite` or `ARS-Codex`. You should **not** see separate `academic-paper`,
  152. `academic-pipeline`, `deep-research`, or `academic-paper-reviewer` skills from
  153. this package. If you do, reinstall with the update command above and open a new
  154. Codex conversation.
  155. ## Codex Docs
  156. - [Codex setup](skills/academic-research-suite/ars/docs/SETUP.md) covers
  157. installation, `ars-*` aliases, optional tools, Material Passport adapters,
  158. and unsupported Claude plugin features.
  159. - [Codex architecture](skills/academic-research-suite/ars/docs/ARCHITECTURE.md)
  160. explains the logical ARS pipeline with the Codex runtime overlay.
  161. - [Optional full-runtime adapter](CODEX_FULL_RUNTIME_ADAPTER.md) documents the
  162. disabled-by-default planner, Codex agent-team templates, and hook pack.
  163. ## Usage
  164. Invoke the suite explicitly with `$academic-research-suite` (singular), then
  165. describe the research task and provide any source files, notes, draft text,
  166. reviewer comments, or output constraints.
  167. ```text
  168. Use $academic-research-suite to help me plan a systematic literature review on
  169. AI adoption in higher education quality assurance.
  170. ```
  171. The Codex adapter routes the request to one of six ARS workflows:
  172. | Workflow | Use when you need | Example prompt |
  173. |---|---|---|
  174. | `deep-research` | Research question refinement, literature review, systematic review, meta-analysis, fact-checking | `Use $academic-research-suite to build a systematic review protocol for AI in higher education QA.` |
  175. | `academic-paper` | Paper outline, drafting, abstract, revision, citation formatting, AI disclosure | `Use $academic-research-suite to turn these notes into an IMRaD paper outline and drafting plan.` |
  176. | `academic-paper-reviewer` | Manuscript review, simulated peer review, editorial decision, re-review | `Use $academic-research-suite to review this manuscript and produce a journal-style decision letter.` |
  177. | `academic-pipeline` | End-to-end research-to-paper workflow with integrity gates, review, revision, and final checks | `Use $academic-research-suite to run an end-to-end research-to-paper pipeline from topic to revised manuscript.` |
  178. | `sr-screener` | Explicitly requested protocol rules, title/abstract or full-text screening, adjudication, audit, and reporting | `Use $academic-research-suite: sr-screener ta-screen these records against my confirmed protocol.` |
  179. | `experiment-agent` | Code experiment planning, human study protocol, statistical interpretation, reproducibility validation | `Use $academic-research-suite to plan a code experiment and define reproducibility checks.` |
  180. `sr-screener` has eight modes: `protocol`, `quick`, `pilot`, `ta-screen`,
  181. `ft-screen`, `adjudicate`, `audit`, and `report`. Screening starts only on an
  182. explicit request; `deep-research` `systematic-review` never hands off to it
  183. automatically. The review team confirms the eligibility rules and verifies AI
  184. decisions. Full title/abstract screening requires a pilot comparison with the
  185. team's labels that misses no record the team advanced, unless the user records
  186. an override. Counts remain provisional until a senior reviewer completes QC
  187. of a reproducible sample of jointly excluded records (minimum 20, default 100; all exclusions when fewer are available).
  188. Missing or malformed decisions remain pending. Synthetic tests exercise the
  189. scripts; screening accuracy has not been measured.
  190. ### Claude-Style Aliases
  191. Claude Code v3.7 installs `/ars-*` slash commands. Codex does not have the same
  192. plugin command registry, so this package emulates the command intent inside the
  193. single `$academic-research-suite` skill. Use either form:
  194. ```text
  195. Use $academic-research-suite: ars-plan my paper on AI governance in universities.
  196. ```
  197. or, when your Codex client passes slash-prefixed text through as a normal user
  198. message:
  199. ```text
  200. /ars-plan my paper on AI governance in universities.
  201. ```
  202. If slash input is intercepted by the client, use the plain alias form:
  203. ```text
  204. ars-plan my paper on AI governance in universities.
  205. ```
  206. | Claude command | Codex alias | Routed workflow |
  207. |---|---|---|
  208. | `/ars-plan` | `ars-plan` | `academic-paper` `plan` mode |
  209. | `/ars-outline` | `ars-outline` | `academic-paper` `outline-only` mode |
  210. | `/ars-abstract` | `ars-abstract` | `academic-paper` `abstract-only` mode |
  211. | `/ars-lit-review` | `ars-lit-review` | `academic-paper` `lit-review` mode |
  212. | `/ars-citation-check` | `ars-citation-check` | `academic-paper` `citation-check` mode |
  213. | `/ars-disclosure` | `ars-disclosure` | `academic-paper` `disclosure` mode |
  214. | `/ars-format-convert` | `ars-format-convert` | `academic-paper` `format-convert` mode |
  215. | `/ars-revision-coach` | `ars-revision-coach` | `academic-paper` `revision-coach` mode |
  216. | `/ars-revision` | `ars-revision` | `academic-paper` `revision` mode |
  217. | `/ars-reviewer` | `ars-reviewer` | `academic-paper-reviewer` full mode |
  218. | `/ars-mark-read` | `ars-mark-read` | Human-read signal for citation keys in the active Material Passport |
  219. | `/ars-unmark-read` | `ars-unmark-read` | Rescind a prior human-read signal |
  220. | `/ars-cache-invalidate` | `ars-cache-invalidate` | Invalidate cached verification entries for one citation key |
  221. | `/ars-full` | `ars-full` | `academic-pipeline` full workflow |
  222. ### Working Pattern
  223. For best results, start with the workflow goal and the current state of your
  224. materials:
  225. ```text
  226. Use $academic-research-suite.
  227. Goal: write a journal article.
  228. Current materials: I have a literature matrix and rough findings, but no outline.
  229. Output needed now: paper architecture and missing-evidence checklist.
  230. Constraints: English, APA 7, higher education policy audience.
  231. ```
  232. If you only have a paper topic or broad research direction and do not yet have a
  233. clear research question, the Codex router should start with ARS Socratic
  234. scoping:
  235. ```text
  236. Use $academic-research-suite.
  237. I want to write a paper on AI adoption in higher education quality assurance.
  238. I do not yet have a clear research question.
  239. Please use SCR / Socratic dialogue to help me narrow the question first; do not write an outline yet.
  240. ```
  241. Expected route: `deep-research` `socratic` mode first. ARS should ask narrowing
  242. questions and should not produce an outline or draft until the research question
  243. has converged.
  244. For review tasks, provide the manuscript or a path to the manuscript, plus the
  245. review mode you want:
  246. ```text
  247. Use $academic-research-suite to review this paper.
  248. Mode: full review.
  249. Focus: methodology, contribution, citation integrity, and likely desk-reject risks.
  250. Output: reviewer reports plus editorial decision letter.
  251. ```
  252. For staged pipelines, ask for a checkpoint instead of asking Codex to run the
  253. entire process silently:
  254. ```text
  255. Use $academic-research-suite to start an academic-pipeline run.
  256. Begin with Stage 0 intake and stop after producing the pipeline dashboard.
  257. ```
  258. ### Smoke Tests
  259. In a new Codex conversation:
  260. ```text
  261. /skills
  262. ```
  263. Expected: one ARS entry only.
  264. Then test Socratic routing:
  265. ```text
  266. Use $academic-research-suite.
  267. I want to write a paper on AI adoption in higher education quality assurance.
  268. I do not yet have a clear research question.
  269. ```
  270. Expected: route to `deep-research` `socratic` mode and ask narrowing questions.
  271. CLI smoke test:
  272. ```bash
  273. codex exec --ephemeral --sandbox read-only \
  274. -C /path/to/academic-research-skills-codex \
  275. 'Use $academic-research-suite. Router smoke test only. User request to classify: I want to write a paper on AI adoption in higher education quality assurance, but I do not yet have a clear research question. According to the academic-research-suite router, classify the workflow and mode.'
  276. ```
  277. Maintainer quality gates:
  278. ```bash
  279. python3 skills/academic-research-suite/codex/scripts/ars_codex_quality_gates.py all --json
  280. ```
  281. Expected: every reported gate has `"ok": true`.
  282. ### Non-Blocking Codex Warnings
  283. These Codex messages do not mean ARS failed to install:
  284. - `[features].codex_hooks is deprecated` — update your Codex config when
  285. convenient; ARS-Codex does not require hooks for normal use.
  286. - `hooks need review before they can run` — review those hooks separately if
  287. you use them. ARS-Codex treats vendored Claude hooks as traceability metadata
  288. and does not require them.
  289. ### Codex Adapter Behavior
  290. ARS was originally written for Claude Code. In this Codex package:
  291. - The vendored `agents/*.md` files are used as role and phase prompts.
  292. - The Codex-only `codex/` directory contains an optional full-runtime adapter
  293. profile. Its fixed topology and hook pack are opt-in; normal execution adapts
  294. native subagents to the task.
  295. - The vendored `commands/ars-*.md` files are prompt recipes only. Codex does not
  296. register them as slash commands.
  297. - The vendored `hooks/hooks.json` file is preserved for upstream traceability
  298. only. Codex does not install Claude Code hooks from this package.
  299. - Codex may delegate bounded independent work through native subagents within
  300. the current task and permissions, while the lead continues useful work.
  301. - Web/source verification uses Codex browsing and must cite sources when current
  302. or external facts matter.
  303. - Cross-model verification is disabled by default. When explicitly requested in
  304. this Codex package, follow the vendored provider setup in
  305. `ars/shared/cross_model_verification.md`, identify the provider/model/content
  306. class first, and obtain explicit user consent before any external upload.
  307. External reviewers are called through configured provider APIs, not simulated
  308. through the active Codex model.
  309. - `ARS_MODEL_TIERING` is unset by default. The Codex adapter preserves the
  310. upstream judgment/execution classification but applies `economy` or
  311. `quality-boost` only when the runtime supports an explicit per-dispatch model
  312. override; otherwise it reports a no-op and keeps the active model.
  313. - Protected top-level agent `tools:` allowlists remain least-privilege role
  314. boundaries. A dispatched checkpoint owner does not gain Bash or network
  315. transport; the dispatching Codex context owns any explicitly consented
  316. cross-model call.
  317. - A `[CROSS-MODEL-HANDOFF v1]` block is a transport request, not a deliverable.
  318. The dispatcher validates it, sends only its payload, applies the mechanical
  319. result routing, and returns judgment work to the original owner.
  320. - In reviewer `full` mode, an explicitly configured and consented cross-model
  321. run swaps the existing Reviewer 2 seat; it never adds a sixth reviewer.
  322. Re-review applies the separate Priority-1 judge pass and records provenance.
  323. Single-family execution and provider fallback are disclosed.
  324. - `ARS_CACHE_STALE_ADVISORY_DAYS` controls the advisory-only cache-age threshold,
  325. while `ARS_CACHE_REVALIDATE=1` opts into live bibliographic re-validation.
  326. These settings apply when the programmatic citation gate is run; stale rows
  327. alone never fail an integrity gate.
  328. - Locally read PDFs run the structural `pdf_read_preflight.py` by default before
  329. page anchors are trusted. `FAIL` and `UNAVAILABLE` remain distinct, and a
  330. missing parser or sidecar is never treated as `PASS`. The v3.20 sandboxed
  331. content classifier is an explicit opt-in advisory and cannot override this
  332. structural preflight.
  333. - `ars-mark-read` requires a user-declared `read_scope` for every new mark.
  334. Explicit unknown and legacy scope-less records remain `coverage_unknown`;
  335. partial coverage stays visible and Codex never infers full-text reading.
  336. - Revision rounds use the v3.20 evidence-bound, non-ranking roadmap and explicit
  337. author-adjudication contract. They also preserve the v3.19-introduced
  338. claim-strength ladder and deterministic numeric, citation, marker, and
  339. protected-term conservation checks as advisory-first guards.
  340. - The upstream v3.18 SessionStart update checker is vendored but not installed
  341. or executed as a Codex hook. Plugin users update with
  342. `codex plugin marketplace upgrade ars-codex` followed by
  343. `codex plugin add ars-codex@ars-codex`; direct skill installs still update by
  344. reinstalling or pulling this repository.
  345. - Upstream references to a "fresh Claude Code session" mean a new Codex
  346. conversation in this package; Material Passport reset semantics still apply.
  347. - If a citation, source, statistic, or journal policy cannot be verified, Codex
  348. should mark it as unverified rather than invent support.
  349. ### ARS v3.24.2 Parity
  350. This package adapts upstream ARS `v3.24.2` at
  351. `fd5ff1b858374219b4d342e38587b66e18f61cbd` wherever Codex has an equivalent
  352. concept, with documented model/runtime overlays.
  353. Bibliographic network behavior is intentionally explicit at the Codex adapter
  354. boundary:
  355. | Research path | Default Codex behavior | Dedicated API/client trigger |
  356. |---|---|---|
  357. | Ordinary topic or candidate discovery | Codex browsing and authoritative web sources | The four Python resolver clients are not launched |
  358. | Prompt-level ingest, deduplication, or source verification | Codex browsing or official metadata pages | “Automatic” lookup wording in vendored prompts does not launch a Python client |
  359. | Script-backed citation-existence gate | Not implied by `ars-full`; Stage 2.5/4.5 still run as integrity checkpoints through the default Codex route | Explicit programmatic-verification request; then Crossref/OpenAlex/Semantic Scholar run for non-manual references and arXiv only when `arxiv_id` exists, subject to cache behavior |
  360. | Claim-standing discovery | Advisory offer after an eligible Stage 2.5/4.5 Claim Registry row | Separate user request plus affirmative plan-bound consent; uses v3.21 discovery adapters, not the single-reference resolver clients |
  361. | Contamination backfill or migration | Never automatic | Explicit migration CLI only |
  362. | Suggested-reference check in review letters | Not run; the attachment says the check did not run | Explicit request; then the four clients run live for the suggested titles and identifiers only |
  363. | Upstream ARS feature | Codex package behavior |
  364. |---|---|
  365. | One installable plugin | Native Codex plugin `ars-codex`, bundling the single `academic-research-suite` skill |
  366. | `/ars-*` slash commands | Emulated as `ars-*` aliases through the skill router; not native slash commands |
  367. | Five upstream skills auto-discovered from `skills/` symlinks | Single Codex router skill selects the workflow and reads the vendored workflow `WORKFLOW.md` files |
  368. | Plugin-shipped agents | Role/phase prompts run inline or as bounded native subagents according to task dependencies and runtime permissions |
  369. | Optional Codex full-runtime profile | Planner, agent-team templates, and hook pack live under `skills/academic-research-suite/codex/`; disabled by default |
  370. | Four commands (`ars-full`, `ars-reviewer`, `ars-revision-coach`, `ars-citation-check`) omit `model:`; light modes retain `model: sonnet` | These commands inherit the current Codex session model; light-mode `sonnet` remains upstream Claude metadata and does not override the session model |
  371. | `ARS_MODEL_TIERING=economy\|quality-boost` | Classification is preserved; routing remains advisory unless Codex exposes per-dispatch model selection |
  372. | Protected agent `tools:` allowlists | Preserved as least-privilege role boundaries; dispatched owners do not receive Bash/network transport |
  373. | Canonical cross-model handoff envelope | Dispatcher validates the envelope, transports only the payload after consent, and follows the closed result-routing contract |
  374. | Contained Codex citation transport | Opt-in, consent-gated transport is limited to narrow citation-integrity checks; it is inactive unless explicitly configured and requested |
  375. | Run ledger and handoff check | With a passport file, pipeline prompts direct the caller to keep a local ledger of exact user words, step receipts, and file hashes; deterministic reports check handoffs after compaction, resume, and subagent returns, without establishing new authorization |
  376. | Deterministic acronym check | The caller checks saved drafts and abstracts locally; partial coverage is explicit, and a review attachment remains advisory and outside decision, roadmap, and re-review criteria |
  377. | Experiment lock validation | An entry may use `repro_lock: null` with a reason, which earlier releases reject; required lock values must meet their field rule, so some passports that passed before now fail |
  378. | Reviewer card confirmation | After Phase 0 the session shows every card and starts no reviewer seat until the user accepts the set in their own turn; only their own request skips it; compliance is unmeasured |
  379. | Review-letter attachments | The suggested-reference check runs only on explicit request; the reviewer claim check uses one isolated, tool-less worker and is never judged inline; neither feeds the decision, roadmap, or re-review criteria |
  380. | Instruction/data boundary and routing | Third-party text remains data across workflow intake, dispatches, and tool reads; explicit requests stay in their selected mode when inputs are missing, and Claude startup hooks remain inactive |
  381. | Chinese APA 7 citation checks | Preserve author-abbreviation exceptions and reference-list authors, require evidence before proposing a stroke-order change, and distinguish visible syntax errors from unverified source claims |
  382. | Evidence-bound review and revision | Durable evidence rows, confirmed review criteria, non-ranking roadmaps, author adjudication, and revision-evidence bundles are preserved |
  383. | Research-workflow profiles | Default-off deterministic selection with a visible field-general fallback; no research family is inferred from manuscript content, and corrections stale rather than rewrite prior artifacts |
  384. | Inquiry branch ledger | `ARS_INQUIRY_LEDGER=1` enables the local opt-in alpha; author events, bounded summaries, path/lock/recovery safeguards, and stale causes are preserved without granting network authority |
  385. | Sealed model-promotion bakeoffs | Commitment/reveal contracts and hermetic tests are vendored; the history-dependent tree verifier remains upstream-only because the re-rooted subtree has no complete upstream Git history |
  386. | Socratic research-question authorship | Non-convergence never triggers system-authored candidate questions; explicit user request is required to leave non-generation mode |
  387. | Categorical reviewer judgement and panel provenance | Live packages remain `NOT_CALIBRATED`; no numeric score, weight, aggregate, ranking, or binary independence claim is fabricated |
  388. | Review criteria and human-subjects authority | Venue/criteria and ethics/data-protection authority require explicit user confirmation; Codex does not infer or simulate approval |
  389. | Optional PDF content classifier | The sandboxed classifier is an opt-in advisory dependency and cannot override structural PDF preflight results |
  390. | Cross-model Reviewer 2 and re-review judge tracks | Available only with explicit provider configuration and content consent; the fixed seat, Judge Record, single-family disclosure, and fallback disclosure are preserved |
  391. | Source-backed review-criteria proving set | Exact-profile author confirmation, source receipts, and three-consumer digest binding are preserved; one proving profile is not venue or discipline coverage |
  392. | Cache staleness advisory and live re-validation | Local cache remains the default; stale rows are advisory-only and `ARS_CACHE_REVALIDATE=1` opts into live bibliographic checks |
  393. | Data-flow and capability transparency | The v3.21 network map, control-availability matrix, stage-capability matrix, risk register, and governance statement are vendored without promoting evidence labels into effectiveness or certification claims |
  394. | Claim-standing pipeline wiring | Eligibility only offers the advisory view; query-plan binding, explicit consent, freshness validation, and transmission accounting remain mandatory before any external call |
  395. | Risk-stratified claim, scope, and novelty checks | Vendored workflow prompts and schemas preserve high-impact-first sampling plus advisory-only scope and search-bounded novelty rows |
  396. | Local-PDF read-integrity preflight | The structural pypdf preflight and sidecar contract remain the default; parser unavailability or repair warnings stay explicit `UNAVAILABLE` advisories, while the v3.20 classifier above remains opt-in only |
  397. | Human-read scope attestation | Every new mark requires user-owned `read_scope`; legacy missing scope remains unknown and partial coverage stays distinguishable from full coverage |
  398. | Claim coverage and bounded evaluation substrates | Exact registered-claim coverage, drift dispositions, claim-standing stance tools, and blind ideation assignment preserve provenance and unmeasured boundaries without certifying semantic completeness or correctness |
  399. | Revision claim-drift guards | The v3.20 non-ranking roadmap and author-adjudication contract complement the claim-strength ladder, revision-evidence bundle, deterministic token-conservation checker, and held-out measurement set |
  400. | Executable panel/degradation/pipeline-boundary checks | Vendored with their hermetic tests and exposed by the optional full-runtime manifest |
  401. | SessionStart and SubagentStop hooks, including the update reminder | Vendored for traceability only; Codex does not install or execute Claude hooks |
  402. | Plugin marketplace update | Refresh with `codex plugin marketplace upgrade ars-codex`, then re-add `ars-codex@ars-codex`; direct skill installs still reinstall or pull |
  403. | Claude Code Agent Team | Native Codex subagents are scheduled adaptively; the separate fixed topology remains opt-in |
  404. | Cross-model provider dispatch from upstream docs | Disabled by default; available only with explicit provider configuration and explicit user consent |
  405. ### Optional External Cross-Model Reviewer API
  406. For reviewer calibration or cross-model devil's advocate checks, configure one
  407. of the provider tuples documented in
  408. `ars/shared/cross_model_verification.md`, then ask for cross-model verification
  409. explicitly in the prompt. For example:
  410. ```bash
  411. export OPENAI_API_KEY="<your-openai-api-key>"
  412. export ARS_CROSS_MODEL="gpt-6-astra"
  413. ```
  414. `gpt-6-astra` remains provisional on both verifier transports. A GPT verifier
  415. used by a GPT session is a separate run in the same family, not cross-family
  416. validation. Provider/content/cost consent still applies.
  417. Without both a configured provider and explicit user consent for the content
  418. class being sent, ARS-Codex falls back to single-runtime review and reports that
  419. cross-model verification was unavailable.
  420. ## Support And Sponsorship
  421. If ARS-Codex helps your research workflow, you can support maintenance through
  422. [Buy Me a Coffee](https://buymeacoffee.com/crucify020v).
  423. ## Security
  424. Do not open public issues for vulnerabilities. Follow
  425. [`SECURITY.md`](SECURITY.md) for private reporting, and see the
  426. [release readiness and security report](security_best_practices_report.md) for
  427. the latest local validation summary.
  428. ### File Layout For Advanced Use
  429. The entry point is:
  430. ```text
  431. skills/academic-research-suite/SKILL.md
  432. ```
  433. Workflow content is under:
  434. ```text
  435. skills/academic-research-suite/ars/<workflow>/
  436. ```
  437. Shared schemas, compliance rules, and cross-workflow contracts are under:
  438. ```text
  439. skills/academic-research-suite/ars/shared/
  440. ```
  441. When debugging or updating the package, preserve these paths. Many ARS workflow
  442. files cross-reference `shared/`, `scripts/`, `examples/`, and other workflow
  443. directories.
  444. ## Update Policy
  445. Updates sync selected upstream ARS content into `skills/academic-research-suite/ars/`.
  446. Do not mirror the Claude Code repo blindly; exclude Claude/plugin loader files
  447. such as `.claude/`, `.claude-plugin/`, source `.gitignore`, and symlink-only
  448. alias directories that are not needed in Codex. Nested upstream `.github/`
  449. workflows may be retained as inactive traceability and self-test fixtures.
  450. ### Inactive Upstream Scripts
  451. Some upstream maintenance scripts are vendored but intentionally inactive in
  452. this Codex package because they require non-vendored Claude Code inputs such as
  453. `.claude/CLAUDE.md`. See `inactive_upstream_scripts` in
  454. `skills/academic-research-suite/manifest.json` before wiring any upstream script
  455. into Codex CI.
  456. ## Contributors And Acknowledgements
  457. **Cheng-I Wu** - Maintainer of the ARS suite and this Codex sibling
  458. distribution.
  459. **Codex** - Assisted with the Codex adapter packaging, router-policy hardening,
  460. test fixes, and release-readiness review under maintainer direction.
  461. **[vinschger](https://github.com/vinschger)** - Reported beginner installation
  462. friction around `python` vs `python3`, which led to clearer setup instructions
  463. for macOS and other environments.
  464. **[Joker2377](https://github.com/Joker2377)** - Helped answer community
  465. installation questions and clarify beginner setup steps in issue discussions.
  466. Vendored upstream ARS contributors are acknowledged in
  467. [`skills/academic-research-suite/ars/README.md`](skills/academic-research-suite/ars/README.md#contributors).

Ratings and comments

Sign in to rate this skill.

Sign in

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