abstract-shotgun
Skill17 starsChanged 3 months agoMITView on GitHub
Generates 4-6 abstract variants with different positioning bets (puzzle-, policy-, theory-, finding-, anomaly-first), tagged by audience and risk, and recommends one for the target journal. Use when a full draft exists and the user wants the abstract written, improved, or reframed.
In this repository38 other files it gives its agents.
- analyzeSkill
- archiveSkill
- carefulSkill
- coauthor-reviewSkill
- codebookSkill
- cover-letterSkill
- data-acquireSkill
- data-cleanSkill
- design-researchSkill
- draft-sectionSkill
- freezeSkill
- guardSkill
- hypothesis-designSkill
- idea-shotgunSkill
- identification-reviewSkill
- journal-fitSkill
- learnSkill
- lit-mapSkill
- llm-checklistSkill
- mstack-initSkill
- mstack-upgradeSkill
- paper-statusSkill
- pdf-ingestSkill
- power-analysisSkill
- preregisterSkill
- r-and-rSkill
- referee-mockSkill
- research-questionSkill
- results-auditSkill
- retroSkill
- robustnessSkill
- scope-challengeSkill
- submission-packSkill
- survey-buildSkill
- theory-buildSkill
- title-shotgunSkill
- unfreezeSkill
- vizSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 1 is distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 12
Stage: write · Voice: writer, anchored to voice.writing_style in .mstack/config.yaml
- Distinctive: line 47
Over the word limit: If the cap is 200, every variant is ≤ 200.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
/mstack:abstract-shotgun
Stage: write · Voice: writer, anchored to voice.writing_style in .mstack/config.yaml
On a full draft. The abstract is the hardest paragraph and the one editors use to desk-reject or send out; generate variants rather than converging on the first.
Procedure
- Load
paper/sections/intro.tex,results.tex,discussion.tex, and.mstack/config.yaml(target journals). - Voice: invoke the
voice.writing_styleskill if set; otherwise clean generic academic prose (short sentences, active verbs, no hedge-stuffing). - Generate 4–6 variants, each leading differently: puzzle-first ("Why does X happen when Y predicts the opposite?"); policy-first ("When governments do X, does it work?"); theory-first ("We argue that X causes Y because Z"); finding-first ("Using a quasi-experiment in X, we find Y"); anomaly-first ("Standard accounts predict X; we find Y"); method-first (methods papers only).
- Each ≤ 250 words (or the journal cap if shorter) and contains the question, the design with its estimand, the data, the finding with a number in substantive units, and the contribution (the abstract bar in
${CLAUDE_PLUGIN_ROOT}/references/writing-conventions.md); no citations, no hedging stack; only the order varies. - Tag each with lead-with style, implicit audience (IPE, comparative, methods, public), and the framing's main risk (e.g. puzzle-first can look unengaged with the literature).
Write the file, per
formatin.mstack/config.yaml. LaTeX:paper/sections/abstract.tex, variants as%comments, the recommended one uncommented inside\begin{abstract} … \end{abstract}. Quarto:abstract.qmdwith<!-- … -->comments.% VARIANT 1 (puzzle-first): % <text> % % VARIANT 2 (policy-first): % <text> ... % --- RECOMMENDED --- <recommended variant text, uncommented>- Recommend the variant that fits the top target journal, justified in one sentence.
Outputs
paper/sections/abstract.tex— variants as comments, one recommended.- Summary block: variants with one-line tags + recommendation.
Anti-patterns
- Variants that open the same way. That is not a shotgun.
- Over the word limit. If the cap is 200, every variant is ≤ 200.
- Ignoring the journal. Some prefer puzzle-first, some finding-first.
Next
/mstack:title-shotgun to pair a title with the chosen framing. /mstack:draft-section abstract gives a baseline beforehand.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

