agentsclimarketplace
Sign in

Short study

Skill MrBridgeHQ/short-study-claude/skills/short-study

Use when turning a simple question into a one-pager data study - a clearly reframed question, a light analysis of a dataset (comparison, correlation, or distribution), and an easy-to-read answer grounded in real, computed figures - plus its promotion assets: anchor links to the study page and social posts for X, Bluesky, and LinkedIn. Triggers on "make a short study", "one-pager study", "is X rated higher than Y", "are A and B correlated", "data study", "study plus social posts", "anchors and posts for a study", EN or FR. Every figure is computed by a script, never estimated.From its SKILL.md

Install
npx -y skills add MrBridgeHQ/short-study-claude --skill short-study

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

5.2 KB, ~1.1k tokens by cl100k_base, as published. Nobody here has run it

Short Study

This skill owns one-pager data studies and their promotion assets. A short study is a single publishable page that takes a vague question, reframes it as a falsifiable claim, backs that claim with computed figures from a real dataset, and presents the answer clearly. Alongside the one-pager, the skill produces anchor links to the study page and social posts for X, Bluesky, and LinkedIn, in EN and FR. Every figure in every output - study or post - is computed by scripts/analyze.py or traced to an explicitly cited external source. Nothing is hand-waved.

Integrity rule (the skill's reason to exist). Every figure that appears in a study or a promotion post must trace to one of two sources: (a) the JSON output of scripts/analyze.py on the supplied dataset, or (b) an explicitly cited external source (URL + retrieval date + the exact quoted number). Hand-computed estimates, approximate figures, and unverifiable statistics are not permitted. If the dataset cannot support a number, the number does not appear.

The pipeline

There is one pipeline: Frame → Analyze → Write → Promote. Most requests start at Frame. The only shortcut: "regenerate posts for an existing study" enters at Promote directly, reading the study's existing figures from the already-published one-pager.

  1. Frame - Translate a vague question into a single falsifiable question of type comparison, correlation, or distribution. Identify the dataset (CSV or JSON file, or a dataset export). Document the reframed question, dataset source, n, and date in analysis.md. Reference: references/analysis.md.

  2. Analyze - Run scripts/analyze.py with the appropriate subcommand (compare, correlate, or distribute) on the dataset file or via stdin (JSON piped in). Capture the full JSON output. All figures used downstream must come from this output. Reference: references/analysis.md.

  3. Write - Produce the one-pager using the fixed structure (reframed question, key finding, chart spec, methodology note, source + date). The chart spec describes axes, series, and type - it is not rendered here. Write EN and FR versions. Reference: references/one-pager.md.

  4. Promote - Produce anchor-text variants for the study URL and social posts (X ≤ 280 chars, Bluesky ≤ 300 chars, LinkedIn long-form), EN and FR. Reference: references/promotion.md.

Available references and script

ResourcePurpose
references/analysis.mdHow to run scripts/analyze.py, interpret its JSON output, document n + source + date, flag sample bias, and what to do when figures are unsupported
references/one-pager.mdFixed structure for the study page - all mandatory sections, EN/FR templates, chart spec format, methodology note, and validation checklist
references/promotion.mdAnchor-text variants + social post templates for X / Bluesky / LinkedIn, EN and FR, with character-count rules
scripts/analyze.pyThe numbers engine - compare, correlate, distribute subcommands; file or stdin; JSON output with n and chart series; stdlib only, no network

Rules

R1 - No live fetching in the script. scripts/analyze.py reads a local file or stdin only - it never opens a network connection. Data must arrive as a pre-downloaded CSV/JSON file or as dataset JSON piped in. The script never invokes requests, httpx, curl, or any HTTP library.

R2 - Integrity and honesty. Every figure must be traceable (see the integrity rule above). Always state n, the dataset source, and the retrieval/export date in the study. Correlation ≠ causation - say so when reporting a correlation. Flag sample bias explicitly if the dataset is non-representative (e.g., only items rated on a single platform, only entries with ≥ 10 reviews).

R3 - Anchor doctrine. Anchor-text selection for the study URL follows standard link-building doctrine: branded + naked-URL anchors dominate; partial-match descriptive anchors fill the middle; exact-match anchors are rare. references/promotion.md shows how to apply these principles to a specific study.

R4 - Human-readable prose. Study prose and all social posts should be reviewed so they read as human-authored before delivery. AI-pattern text in a published study undermines credibility.

What this skill is not

  • Not the anchor-text doctrine itself. This skill applies common link-building principles; it does not own or re-derive a full off-page SEO strategy.
  • Not a chart renderer. The skill emits a chart specification (axes, series, type) embedded in the one-pager. Rendering the chart as an image is outside scope.

What ships with it: 8 files

58.3 KB alongside SKILL.md, 1 of them executable

assets/

references/

scripts/

Gives 0 of the 12 instructions most learn study skills give in ~1.1k tokens

Counted across 546 of the 573 authors here whose files we hold, read 2026-08-07

  • Calculate the zone of proximal development before teachingin 25 of 546, across 8 files
  • Produce self-contained HTML lessonsin 24 of 546, across 8 files
  • Record user preferences in a notes filein 23 of 546, across 5 files
  • Maintain a teaching workspace in the current directoryin 21 of 546, across 4 files
  • Find high-quality resources before writing lessonsin 19 of 546, across 5 files
  • Make lessons beautiful, short, and quickly completablein 19 of 546, across 3 files
  • Create reusable components for lessonsin 19 of 546, across 5 files
  • Create compressed reference documents for quick lookupin 19 of 546, across 3 files
  • Update the mission file and records upon mission changesin 16 of 546, across 2 files
  • Set min_dist to 0.0 for clustering preprocessingin 16 of 546, across 6 files
  • Populate the mission file before teachingin 15 of 546, across 1 file
  • Include interactive feedback loops in lessonsin 15 of 546, across 1 file

Said here and by no other author read

  • run the analysis script on the supplied dataset
  • capture the full JSON output from the analysis
  • produce English and French one-pager versions
  • produce English and French social posts
  • state the sample size, source, and date
  • state that correlation is not causation

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.

Keep looking

Skills are one crate of 326,144. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.