expecon-topic-selection — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited expecon-topic-selection (Agent Skill) and scored it 96/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 1 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 1 flagged
The text {match} tells the agent to skip the normal "ask the user first" gate. Used adversarially it removes the human-in-the-loop check before destructive or sensitive actions, turning a normally-gated agent into a fire-and-forget executor.
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Experimental Economics is the ESA method flagship. The editors do not ask "is this topic important?" so much as "is this the cleanest experiment that could answer this question, and does the design itself teach the field something?" Three questions decide fit:
The center of an ExpEcon paper is the minimal pair: two conditions identical except for the one thing your hypothesis is about. Spend your design budget here.
| Venue | What it rewards | Send there instead when… |
|---|---|---|
| JEBO | broad behavioral/organizational questions, method-agnostic | the topic, not the design, is the contribution; you need deception |
| GEB | game-theoretic theory | the experiment merely illustrates a theorem |
| AEJ: Micro | applied micro where an experiment supports a wider claim | the headline is an economic phenomenon, the experiment one leg |
| JESA | short-format ESA work: replications, null results, software, comments | the paper is a brief note, not a full design paper |
The same question can be a lab, lab-in-the-field, or field experiment, and the choice fixes your later bottlenecks:
| Type | Buys you | Costs you | Choose when |
|---|---|---|---|
| Lab | maximal control, clean minimal pairs, cheap iteration | external validity questions (often student pool) | the mechanism needs tight control to isolate |
| Lab-in-the-field | a relevant population with much control retained | recruitment/logistics, some control loss | the population is the point (farmers, traders, CEOs) |
| Field | behavior in situ, strong external validity | partial control, attrition, often costlier | the real-world stakes are the contribution |
Lab control is your comparative advantage at this journal; do not give it up unless the population or the in-situ behavior is the contribution.
A useful litmus for ExpEcon fit: imagine your treatment effect comes out exactly zero. Is the paper still publishable? If yes — because the design is decisive, the hypotheses were pre-registered, and a null adjudicates between behavioral models — you have a true method-defined contribution. If a null would be unpublishable because the paper rests entirely on getting a "surprising positive," you are leaning on the result, not the design, and you risk a publication-bias incentive to p-hack. The flagship (and especially the Registered Report track) values designs whose answer matters either way. Build the question so the null is informative.
A team has "a paper on whether social media reduces cooperation." Topic-first, and unidentifiable as stated. Reframing method-first: design a repeated public-goods game where one treatment injects a between-round "feed" of (real, not fabricated — gate!) peer messages and the control does not. Now the contribution is a treatment effect of peer messaging on contributions, the minimal pair manipulates only the feed, incentives are real, no deception is needed (messages are genuine), and a pre-registered primary comparison (mean contribution, Feed vs. NoFeed, at the matching-group level) exists before any data. The vague topic became a clean ExpEcon design.
【Journal】Experimental Economics (ESA method flagship)
【Skill】expecon-topic-selection
【Verdict】fit / reframe / reroute
【The contribution】treatment effect or design/method advance (one sentence)
【Experiment type】lab / lab-in-field / field + reason
【Minimal-pair contrast】the single manipulated dimension
【Gate check】incentives salient? no deception? [Y/N]
【Pre-reg status】PAP / Registered Report / replication / none-yet
【Why here not sibling】JEBO/GEB/AEJ:Micro/JESA reason
【Next skill】expecon-literature-positioning~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.