New: 98,000+ skills indexed from GitHub repos with 100+ stars · every skill safety-checkedBrowse top skillsTop skills
Esc to closeBrowse all skills
Submit a skillBrowse skills
Submit a skillBrowse skills

Agent skill

Replication Package

by flonat in flonat/flonat-research

Assemble, anonymize, validate, or audit a research replication package.

MITAuto-check passedResearch & Science

Install Replication Package

skills CLI
$ npx skills add flonat/flonat-research --skill replication-package -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install flonat/flonat-research replication-package --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/flonat/flonat-research.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/replication-package .claude/skills/replication-package && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
replication-package
GitHub stars
145
Token cost
~2.5k tokens
SKILL.md length
1,062 words
Files
14 (incl. references)
Skills in repo
83
Repo updated
First seen
Licence
MIT

At a glance

Assemble, anonymize, validate, or audit a research replication package.

  • Works in 7 steps: Never modify the original project. All… → Dry-run mandatory. Always show what will… → Binary files are never modified. Warn… → …
  • Preparing code and permitted data for reviewer
  • SKILL.md covers Modes, When to Use, When NOT to Use and Critical Rules, plus 7 more sections
  • Calls uv, rsync and git

What it does

Replication Package is an agent skill from flonat/flonat-research. Assemble, anonymize, validate, or audit a research replication package. Use when preparing code and permitted data for reviewer or public release. Not for auditing code quality alone; use $code-suite or $replication-audit as appropriate.

Its SKILL.md is about 2.5k tokens, which your agent loads only when the skill is triggered. The skill folder holds 14 other files, including reference files (for example `references/aea-readme-template.md`, `references/anonymization-patterns.md` and `references/assemble-workflow.md`).

It sits in Research & Science, covering Econometrics and empirical research. The repository describes itself as: Shareable Claude Code + Codex infrastructure for PhD researchers — skills, agents, hooks, and rules for academic workflows. The licence is MIT.

When your agent uses it

  • Preparing code and permitted data for reviewer
  • Tasks that involve Econometrics and empirical research

Example prompts

  • “/replication-package”

Requirements

  • Python 3
  • Pre-approved tools (allowed-tools): Bash(cp*), Bash(rm*), Bash(mkdir*), Bash(ls*), Bash(git*), Bash(find*), Bash(sed*), Bash(grep*), Bash(du*), Bash(wc*), Bash(dirname*), Bash(basename*), Bash(readlink*), Bash(rsync*), Bash(mv*), Read, Write, Edit, Glob, Grep, AskUserQuestion, Skill

Workflow steps

7 steps, taken from the first numbered list in SKILL.md.

  1. Never modify the original project. All work happens on a copy in a sibling folder.
  2. Dry-run mandatory. Always show what will be removed/changed and get user confirmation before any deletions.
  3. Binary files are never modified. Warn the user to manually check PDFs (Document Properties), images (EXIF data), and datasets for embedded…
  4. Self-citations are always interactive. Never auto-remove or auto-anonymize a citation. Flag each potential self-citation and let the user…
  5. Resolve symlinks. Use rsync -aL so symlinked content (e.g., Overleaf paper/ symlinks) becomes real files in the copy.
  6. Preserve compilability. The output must still compile/run — only infrastructure and identity are removed, not project functionality.
  7. Blind mode runs the structured-metadata field check (A2) before reporting clean. pyproject.toml [project] authors, package.json…

What it can do on your machine

Read from SKILL.md and the folder at commit da27600. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves these tools, so the agent can use them without asking each time:

    • Bash(cp*)
    • Bash(rm*)
    • Bash(mkdir*)
    • Bash(ls*)
    • Bash(git*)
    • Bash(find*)
    • Bash(sed*)
    • Bash(grep*)
    • Bash(du*)
    • Bash(wc*)

    …and 12 more on the same allowed-tools line.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Shell commands in SKILL.md call:

    • uv
    • rsync
    • git

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md. Its commands use uv, rsync and git, which can reach the network depending on how they are called.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Replication Package loads about 2.5k tokens when it runs, and up to ~19k if it reads all its reference files. Until then it costs about 64 tokens; SKILL.md has 1,062 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~64
When it runs · the whole SKILL.md, loaded when a task matches
~2.5k
With references · SKILL.md plus every file in references/, read only if the agent opens them
~19k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from flonat/flonat-research at commit da27600, republished under its MIT licence (© flonat). 1,062 words, ~2,509 tokens.

Download SKILL.mdSave it as .claude/skills/replication-package/SKILL.md (or your agent's skills folder). This skill also uses 13 other files; get the full folder from GitHub.
name
replication-package
description
Assemble, anonymize, validate, or audit a research replication package. Use when preparing code and permitted data for reviewer or public release. Not for auditing code quality alone; use $code-suite or $replication-audit as appropriate.
allowed-tools
Bash(cp*), Bash(rm*), Bash(mkdir*), Bash(ls*), Bash(git*), Bash(find*), Bash(sed*), Bash(grep*), Bash(du*), Bash(wc*), Bash(dirname*), Bash(basename*), Bash(readlink*), Bash(rsync*), Bash(mv*), Read, Write, Edit, Glob, Grep, AskUserQuestion, Skill
argument-hint
[project-path] [--mode assemble|blind|audit]
skill-dependencies
latex

Replication Package — Assemble, Anonymize, and Audit

Build publication-ready replication packages, optionally anonymize for double-blind review, or audit an existing package for reproducibility. The original project is never modified.

Modes

ModeWhat it doesUse case
AssembleClean copy + AI trace removal + AEA-style READMESharing, Zenodo deposit, journal supplementary
BlindEverything in Assemble + identity anonymizationDouble-blind conference/journal submission
AuditRead-only 11-check reproducibility validationPre-deposit quality gate, self-check

Default mode: Assemble. Infer Blind if the user says "anonymize", "double-blind", or "blind review". Infer Audit if the user says "audit", "check reproducibility", or "validate package".

When to Use

  • Submitting supplementary materials or replication files to a journal
  • Depositing a package on Zenodo, Dataverse, or ICPSR
  • Sharing a project repo publicly (GitHub, institutional repository)
  • Preparing for double-blind submission (Blind mode)
  • Self-checking reproducibility before deposit (Audit mode)

When NOT to Use

  • Quick one-off cleanup of a single file — do it manually
  • Removing a single AI artifact — just delete it directly
  • Projects with no empirical or computational component

Critical Rules

  1. Never modify the original project. All work happens on a copy in a sibling folder.
  2. Dry-run mandatory. Always show what will be removed/changed and get user confirmation before any deletions.
  3. Binary files are never modified. Warn the user to manually check PDFs (Document Properties), images (EXIF data), and datasets for embedded metadata.
  4. Self-citations are always interactive. Never auto-remove or auto-anonymize a citation. Flag each potential self-citation and let the user decide per citation. In Blind mode, surface every self-citation against the loaded submission author list and require a per-citation decision (third-person OK / blind the entry). When the cited paper's author list is a subset of the submission's, third-person is structurally insufficient and the bib entry MUST be blinded — see _shared/double-blind-anonymity-checklist.md §P4–P5.
  5. Resolve symlinks. Use rsync -aL so symlinked content (e.g., Overleaf paper/ symlinks) becomes real files in the copy.
  6. Preserve compilability. The output must still compile/run — only infrastructure and identity are removed, not project functionality.
  7. Blind mode runs the structured-metadata field check (A2) before reporting clean. pyproject.toml [project] authors, package.json author/contributors, Cargo.toml [package] authors, CITATION.cff, LICENSE holder, etc. — see _shared/double-blind-anonymity-checklist.md §"Structured-metadata field check (A2)" for the full target list. This was the CCS 2026 #1328 desk-reject trigger and is now non-skippable.

Assemble Mode (Non-Blind)

Phases 1-7: Scan → Copy → Scrub AI Traces → Generate README → Verify → Fresh Git → Report.

Full workflow: references/assemble-workflow.md

Blind Mode (Assemble + Anonymization)

Runs all Assemble phases, then continues with Phases 8-12: Collect Identity → Anonymize LaTeX → Anonymize Other Files → Anonymous Git → Leak Check Report.

Full workflow: references/blind-workflow.md

Audit Mode (Read-Only)

11-check reproducibility validation: Compilation, Script order, Output presence, Dependencies, Data provenance, README, File sizes, End-to-end clarity, AI traces, Identity leaks, Numeric reproduction.

Full workflow: references/audit-workflow.md

Check 11 (Numeric reproduction) is N/A unless an expected_values.json (the manuscript's reported numbers) sits at the package root. When present, the audit parses the package's committed output files and scores them against that ground truth within tolerance — catching paper-vs-output drift. It is read-only and never re-runs scripts. Convention + scoring rubric: references/expected-values-schema.md.

HPC-run results

If the project used [HPC cluster] (hpc/ directory with *.sbatch), the results in out/<jobid>/ should include git-sha.txt + git-status.txt (written by the sbatch templates before srun). Audit must verify these exist and the SHA matches a commit in the repo — this is the compute-reproducibility equivalent of Script order + Dependencies for HPC runs. The Assemble README should document the hpc/ entry point and the HF/conda env-setup script alongside code/ + data/. See Task Management docs/guides/hpc.md.

Show full SKILL.md (499 more words)Show less

What This Skill Does NOT Do

  • Does not modify the original project — all changes are in the sibling folder (Assemble/Blind) or purely read-only (Audit).
  • Does not modify binary files — PDFs, images, datasets are copied as-is. User must check these manually for embedded metadata.
  • Does not auto-handle self-citations — every potential self-citation requires user decision.
  • Does not anonymize the paper title — titles are generally not considered identifying (but some venues disagree; user should check).
  • Does not strip PDF metadata — if a compiled PDF exists, its Document Properties may contain author info. User should recompile from the anonymized source or use exiftool to strip metadata.
  • Does not run scripts — it verifies their presence and order but does not execute them (too risky without a controlled environment). Check 11 (Numeric reproduction) likewise compares the package's already-committed output files against the manuscript; it never re-runs the pipeline, so it catches paper-vs-output drift, not full fresh-run reproduction.

Examples

Assemble mode (default)

"Build a replication package for my research paper"

Runs Assemble mode on the current project, creates ../mcdm-paper-replication/.

Blind mode

"Anonymize my paper for AAAI double-blind submission"

Runs Blind mode, creates ../mcdm-paper-replication-blind/.

Audit mode

"Audit the reproducibility of my replication package"

Runs Audit mode (read-only) on the specified package directory.

Explicit path and mode

"replication-package <project-path> --mode blind"

Runs Blind mode on the specified project path.

Cross-References

  • data-sensitivity rule — raw data is read-only; replication packages must document data provenance without modifying data/raw/
  • overleaf-separation rule — paper/ structure is preserved; symlinks resolved by rsync
  • shared/multi-language-conventions.md — dependency detection patterns for Python, R, Julia, MATLAB
  • shared/publication-output.md — output file verification and freshness checks
  • latex — compilation check in Audit mode
  • references/aea-readme-template.md — AEA-style README template for Assemble mode
  • references/figure-table-crosswalk.md — per-figure/table crosswalk (with LaTeX Label) + paper-consistency check, appended to the README (Phase 4)
  • references/logging-skeletons.md — per-script logging + master-script skeletons (R/Python/Julia/Stata) offered in Phase 5
  • references/release-readiness-checklist.md — 14-point PASS/FAIL pre-release gate emitted at Phase 7 (RELEASE-READINESS.md)
  • references/rules.dropboxignore — Dropbox-sync ignore file for Dropbox-synced packages (Phase 6)
  • references/anonymization-patterns.md — replacement tables for Blind mode
  • references/audit-rubric.md — 11-check scoring rubric for Audit mode
  • references/expected-values-schema.md — expected_values.json convention + numeric-reproduction scoring rubric (Check 11)
  • references/deposit-checklist.md — platform-specific deposit completeness checklist
  • references/report-template.md — report format for Blind mode
  • _shared/double-blind-anonymity-checklist.md — authoritative paper+artifact anonymity matrix (P1–P8, A1–A9). Blind mode must run all artifact-side checks (A1–A9) before reporting clean.

Note: This skill replaces the former /export-project-clean and /export-project-anon skills. All their functionality is preserved in Assemble and Blind modes respectively.

Output Verification (Guard)

This skill writes files. Before any auto-commit, emit an outputs manifest and run the shared verifier. See skills/_shared/verify-outputs.md for the full protocol.

Required tail steps (before git commit):

  1. Write the manifest to <project>/.context/state/outputs-manifest-<UTC-timestamp>.json, listing every file this skill claims to have written in this invocation (paths relative to the project root).

  2. Run:

    bash
    uv run python "<skills-root>/_shared/verify_outputs.py" \
    --manifest "$MANIFEST" \
    --project-root "$PROJECT_ROOT"

  3. If the verifier exits non-zero, do not commit — surface the missing-files list to the user and stop. The verifier has already logged an error entry to ~/.local/state/ai-workflows/skill-outcomes.jsonl, which feeds the shared skill-health dashboard.

Why: closes the "hallucinated outputs" failure class (commit b2cff75, 2026-04-18).

© flonat, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 13 other files (references) in skills/replication-package of flonat/flonat-research.

  • SKILL.md
  • references/aea-readme-template.md
  • references/anonymization-patterns.md
  • references/assemble-workflow.md
  • references/audit-rubric.md
  • references/audit-workflow.md
  • references/blind-workflow.md
  • references/deposit-checklist.md
  • references/expected-values-schema.md
  • references/figure-table-crosswalk.md
  • references/logging-skeletons.md
  • references/release-readiness-checklist.md
  • references/report-template.md
  • references/rules.dropboxignore

Open the folder on GitHubat commit da27600

Compare with similar skills

Replication Package next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Replication Package compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Replication Package this skillflonat/flonat-research145—~2.5kAutomated safety check: PassMIT
Diagnosepedrohcgs/claude-code-my-workflow1.6k—~4.2kAutomated safety check: PassMIT
Aer Replicationbrycewang-stanford/Auto-Empirical-Research-Skills4.5k1 repos~2.8kAutomated safety check: PassCustom licence
New Analysisbrycewang-stanford/Auto-Empirical-Research-Skills4.5k—~764Automated safety check: NotesCustom licence
Eer Replication Packagebrycewang-stanford/Awesome-Journal-Skills1.2k—~1.2kAutomated safety check: PassMIT
Statadylantmoore/stata-skill2911 repos~4.2kAutomated safety check: PassCustom licence

Similar skills

  • Diagnose

    pedrohcgs/claude-code-my-workflow

    Root-cause a failing or wrong empirical result with a disciplined reproduce → minimise → hypothesise → instrument → fix loop, instead of guessing-and-poking.

    1.6k GitHub stars~4.2k tokensUpdated 10 days ago
    Research & ScienceAuto-check passed

  • Aer Replication

    brycewang-stanford/Auto-Empirical-Research-Skills

    A skill your agent uses when assembling the AEA Data and Code Availability deposit for an AER, AER:Insights, or AEJ acceptance, writing the README, or auditing a replication package before the AEA…

    4.5k GitHub starsUsed in 1 repo~2.8k tokens
    Research & ScienceAuto-check passed

  • New Analysis

    brycewang-stanford/Auto-Empirical-Research-Skills

    Scaffolds a method-specific analysis notebook (DiD, IV, RDD, LASSO, Panel FE) with boilerplate.

    4.5k GitHub stars~764 tokensUpdated 2 days ago
    Research & ScienceAuto-check: notes

  • Eer Replication Package

    brycewang-stanford/Awesome-Journal-Skills

    A skill your agent uses when assembling the data and code deposit for a European Economic Review (EER) manuscript under Elsevier's mandatory replication policy.

    1.2k GitHub stars~1.2k tokensUpdated 11 days ago
    Research & ScienceAuto-check passed

  • Stata

    dylantmoore/stata-skill

    Comprehensive Stata reference for writing correct .do files, data management, econometrics, causal inference, graphics, Mata programming, and 20 community packages (reghdfe, estout, did, rdrobust…

    291 GitHub starsUsed in 1 repo~4.2k tokens
    Research & ScienceAuto-check passed

  • Stata C Plugins

    dylantmoore/stata-skill

    Develop high-performance C/C++ plugins for Stata using the stplugin.h SDK.

    291 GitHub starsUsed in 1 repo~5.8k tokens
    Research & ScienceAuto-check passed

More from flonat/flonat-research

All 83 skills in this repo

  • Latex Posters

    flonat/flonat-research

    Create a large-format academic poster in LaTeX using beamerposter, tikzposter, or baposter.

    145 GitHub stars~1.5k tokensUpdated 8 days ago
    Auto-check: notes

  • Skill Creator

    flonat/flonat-research

    Create, revise, and evaluate reusable AI workflow skills, including trigger-quality tests.

    145 GitHub stars~4.4k tokensUpdated 8 days ago
    Auto-check passed

  • DOCX

    flonat/flonat-research

    Create, read, edit, or convert Microsoft Word documents while preserving professional document structure.

    145 GitHub stars~1.2k tokensUpdated 8 days ago
    Auto-check passed

  • PDF

    flonat/flonat-research

    Read, create, combine, split, rotate, OCR, watermark, secure, or extract content from PDF files.

    145 GitHub stars~488 tokensUpdated 8 days ago
    Auto-check passed

  • Init Project Orchestration

    flonat/flonat-research

    Create or migrate project-level agents, repeatable project workflows, and planning state from one client-neutral contract, then render repository-scoped adapters for both Claude Code and Codex.

    145 GitHub stars~1.6k tokensUpdated 8 days ago
    Auto-check passed

  • Pre Commit Audit

    flonat/flonat-research

    Deliver a fast pre-commit safety scan: file size, anonymity (author / affiliation strings in tex/bib), hardcoded secrets, and invisible-Unicode carriers.

    145 GitHub stars~2.8k tokensUpdated 8 days ago
    Auto-check: notes

Categories

Questions about Replication Package

What does Replication Package do?

Assemble, anonymize, validate, or audit a research replication package. Replication Package is an agent skill from flonat/flonat-research. Assemble, anonymize, validate, or audit a research replication package.

When should I use Replication Package?

Replication Package fits situations like: preparing code and permitted data for reviewer; tasks that involve Econometrics and empirical research.

How do I install Replication Package in Claude Code?

Run `npx skills add flonat/flonat-research --skill replication-package -a claude-code`. Or copy the skill folder (skills/replication-package in flonat/flonat-research) into .claude/skills/replication-package in your project. Claude Code loads it when a task matches its description.

How do I install Replication Package in Codex?

Run `npx skills add flonat/flonat-research --skill replication-package -a codex`. Or copy the skill folder (skills/replication-package in flonat/flonat-research) into .agents/skills/replication-package in your project. Codex loads it when a task matches its description.

Can I use Replication Package in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add flonat/flonat-research --skill replication-package -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/replication-package, .gemini/skills/replication-package, .github/skills/replication-package and .opencode/skills/replication-package in your project.

What does Replication Package need to run?

Going by SKILL.md and its folder, Replication Package needs the command-line tools its instructions call (uv, rsync and git). Our summary lists: Python 3. Its frontmatter pre-approves these tools: Bash(cp*), Bash(rm*), Bash(mkdir*), Bash(ls*), Bash(git*), Bash(find*), Bash(sed*), Bash(grep*), Bash(du*), Bash(wc*), Bash(dirname*), Bash(basename*), Bash(readlink*), Bash(rsync*), Bash(mv*), Read, Write, Edit, Glob, Grep, AskUserQuestion, Skill.

Does Replication Package access the network?

SKILL.md contains no URLs. Its commands use uv and git, which can reach the network depending on how they are called. This is read from the text; nothing was executed.

Is Replication Package safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Replication Package use?

Replication Package is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Replication Package use?

About 2.5k tokens (SKILL.md is roughly 10k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 16k tokens, read only when the agent opens those files.

What are the alternatives to Replication Package?

Skills that share tags, products or a category with Replication Package: Diagnose (pedrohcgs/claude-code-my-workflow, 1.6k stars), Aer Replication (brycewang-stanford/Auto-Empirical-Research-Skills, 4.5k stars), New Analysis (brycewang-stanford/Auto-Empirical-Research-Skills, 4.5k stars) and Eer Replication Package (brycewang-stanford/Awesome-Journal-Skills, 1.2k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Replication Package?

flonat (a GitHub user) maintains it in flonat/flonat-research, which has 145 GitHub stars. The repository holds 83 skills in this directory. The repository was last updated on September 29, 2026.

Source: flonat/flonat-research on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.