Run a study the preregistered way with a tamper-evident record: hypotheses, protocol frozen by hash, decisions, deviations, gates, guarded expensive runs, budget, confirmatory vs exploratory results.
Install with the Skills CLI (verifies the checksum, previews the destination, never runs scripts)
skills install @spboucher/preregistered-research@1.0.0 --client claude-codeAdd --project to install into .claude/skills/preregistered-research instead of ~/.claude/skills/preregistered-research.
1Or manually: download the release archive
curl -fsSLo preregistered-research-1.0.0.tar.gz https://skills.spboucher.ai/api/v1/artifacts/spboucher/preregistered-research/versions/1.0.0/download2Verify the SHA-256 checksum
echo "1c6e4c13b6a95e218523d483519bf48adc6170df3d5e8a2bb691200c09e00973 preregistered-research-1.0.0.tar.gz" | shasum -a 256 -c -3Inspect the contents before extracting
tar -tzf preregistered-research-1.0.0.tar.gz4Extract into ~/.claude/skills/preregistered-research
mkdir -p ~/.claude/skills/preregistered-research && tar -xzf preregistered-research-1.0.0.tar.gz -C ~/.claude/skills/preregistered-researchInstalling only copies or configures. Nothing is executed for you. The Skills CLI verifies the checksum before writing files.
An agent skill for running an empirical study the preregistered way, with a tamper-evident record.
It teaches the method (design review before data, a frozen protocol, numbered decisions and deviations, gates, guarded expensive steps, a budget, honest reporting of null results) and ships a small command-line tool that keeps the record: scripts/ledger.mjs.
| Path | Content |
|---|---|
SKILL.md | The sequence and the rules |
references/method.md | The method in full |
references/templates.md | Protocol, decision, deviation, gate review, transparency report |
references/cli.md | Every command, exit codes, file layout, status rules |
scripts/ledger.mjs | The command-line tool |
scripts/lib/ledger.mjs | The engine, identical to the one in research-ledger-mcp |
Node.js 20 or later. No packages to install, no network access, no account.
cd my-study
node /path/to/skill/scripts/ledger.mjs init --title "My study" --budget 40
node /path/to/skill/scripts/ledger.mjs hypothesis --statement "x raises y in the full sample" --test "OLS of y on x, clustered by unit" --rule "coefficient > 0, one-sided p < 0.05"
node /path/to/skill/scripts/ledger.mjs freeze --file protocol.md --file analysis.py
node /path/to/skill/scripts/ledger.mjs status
The record is written to .research/ in the project. Commit it with the work.
Reads the files you list (to hash them) and writes inside .research/ of the project root. Runs with node, which is why the manifest declares shell.
MIT
Select a file to preview it. Files are shown as plain text and never executed.
No advisory findings from static-1.0.0 on Nov 10, 2026. 3 informational notes.
Static checks are advisory. They never certify that a package is safe. Compatibility entries are publisher claims unless marked verified.
| Version | Published | Size | SHA-256 | State |
|---|---|---|---|---|
| 1.0.0latest | Nov 10, 2026 | 31 kB | 1c6e4c13b6a9… | published |
Run a sequential revision experiment on a text and record what the Pangram detector reports for each version, as an observed outcome: frozen editorial plan, one submission per step, real evaluations, preservation checks, chart and report.
Peer-review protocol for LaTeX and Markdown manuscripts: deterministic audit script, five review roles, adversarial verification of findings, anchored rubric, referee report and decision letter templates.
Build, audit and grade Excel model workbooks: layout standard, dependency-free command-line auditor (errors, hard-coded numbers, broken patterns, circular references, controls, stale values) and rubric-based grading aid.