Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add mimaowang/econ-writing-skills --skill write-economics-conclusiongit clone --depth 1 https://github.com/mimaowang/econ-writing-skillsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/mimaowang/econ-writing-skills/write-economics-conclusion)<a href="https://agentmods.dev/skills/mimaowang/econ-writing-skills/write-economics-conclusion"><img src="https://agentmods.dev/badge/skills/mimaowang/econ-writing-skills/write-economics-conclusion/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/mimaowang/econ-writing-skills/write-economics-conclusion"><img src="https://agentmods.dev/badge/skills/mimaowang/econ-writing-skills/write-economics-conclusion.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00117 | $0.02706 |
| Opus 5 | $0.00059 | $0.01353 |
| Sonnet 5 | $0.00023 | $0.00541 |
| Haiku 4.5 | $0.00012 | $0.00271 |
Grade A, and why
write-economics-conclusion scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 12d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 156 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Write an Economics Conclusion
Create the clean snapshot a reader should carry away from the paper: what the paper established, why it matters, and where the claim stops. Synthesize rather than repeat. Treat every extension beyond the results—mechanism, external validity, policy, and future work—as evidence-dependent.
Load the required knowledge
Before drafting, rewriting, or auditing, read all four references:
references/evidence-protocol.md— manuscript reading, evidence ledger, causal language, conflicts, and missing information.references/conclusion-knowledge.md— the complete guide-derived principles, rationales, conditions, and anti-patterns.references/composition-and-length.md— module selection, natural order, transitions, paper-type branches, and elastic length.references/quality-and-sources.md— final audit, scoring rules, Top Five calibration, and provenance.
These references are mandatory. Follow the target journal and any separate Discussion-section convention before applying default length guidance.
Non-negotiable rules
- Base the conclusion on claims delivered by the manuscript body, not merely promised in the introduction.
- Never add a result, coefficient, mechanism, limitation, citation, counterfactual, future project, or policy claim that the supplied material does not support.
- Distinguish identified mechanisms, model-implied mechanisms, evidence consistent with a channel, suggestive evidence, and author speculation.
- Preserve mechanism strength exactly. If the source is only
consistent with, supportive, or suggestive, do not writeprincipal/main channel,drives,explains,accounts for,operates through, or a causalbecauseunless separately supported. - Map every explanatory clause, appositive, modifier, adjective, and comparison—not only the sentence's main result. A correct main clause does not license an unsupported
where,because,especially when, orreflectingexplanation. - Match causal language to the design and external-validity language to the studied population, period, and institution.
- Use only the few numbers needed to preserve economic magnitude or a policy tradeoff; verify every component.
- Treat a number's denominator, baseline, comparison, and normalization as separate facts. Never infer
of baselineor another denominator from nearby context. - Preserve every uncertainty operator on a null result.
No detectable effectmay not becomeno effect,unchanged,unaffected,without reducing/eroding/harming,preserved, ordid not harm; every restatement must carry its own detection and proxy qualifications. - Treat
low-cost,large,substantial,modest,meaningful,efficient, and similar evaluations as factual comparisons. Use them only with a reported benchmark or explicit source characterization. - Apply a closed-world paraphrase rule: a smoother expression is eligible only if the source entails it without adding scope, causal force, mechanism status, measurement meaning, evaluation, or centrality.
- Do not convert a robustness exercise into a contribution or a caveat into proof that the concern is resolved.
- Keep the formal conclusion in English by default. Use the user's language for necessary questions or risk notes.
- Do not modify the source manuscript unless explicitly asked.
What ships with it
11 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- agents/openai.yaml 136 B
- evals/evals.json 2.5 KB
- evals/files/observational-remote-work-paper.md 1.5 KB
- evals/files/rct-transit-paper.md 1.8 KB
- evals/files/theory-ai-delegation-paper.md 1.4 KB
- evals/trigger-eval.json 986 B
- references/composition-and-length.md 6.9 KB
- references/conclusion-knowledge.md 8.5 KB
- references/evidence-protocol.md 14 KB
- references/quality-and-sources.md 7.0 KB
- scripts/validate_output.py 2.5 KB runs code
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 12d ago First seen · 156 lines · 117 tokens per session scan A 77049a12d33b
write-economics-conclusion is a skill published in the GitHub repository mimaowang/econ-writing-skills (6 stars, last pushed 1mo ago), licensed MIT. It adds 117 tokens to every session and 2,706 once invoked, about $0.0006 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
paper-reviewer
Deep pre-submission review of a scientific manuscript, modeled on Google's Paper Assistant Tool (PAT). Segments the manuscript, allocates a reasoning budget per segment, dispatches deep reviewers in parallel (each with the full text as context), and consolidates into a single report with severity, quoted evidence, and…
paper-writer
Segmented drafting of a scientific manuscript, with mandatory grounding in declared project artifacts and a project-defined authorial voice. Segments what remains to be written, allocates a reasoning budget per segment, dispatches writers in parallel (each with the full manuscript as context), and consolidates into a…
review-paper
Comprehensive manuscript review with three modes: single-pass (default), --adversarial critic-fixer loop, and --peer [journal] simulated peer-review pipeline (editor + 2 dispositioned referees + editorial decision, calibrated to a target journal). R&R continuation via --peer --r2/--r3; hostile-editor stress test via…
audit-reproducibility
Enforce the replication-protocol.md rule by cross-checking numeric claims in a manuscript against the actual R / Stata / Python outputs. Report PASS/FAIL per claim against tolerance thresholds. Use before submission and before releasing a replication package.
capture-environment
Snapshot the computational environment for a replication package — detects the analysis stack (R / Stata / Python) and emits the right lockfiles (renv.lock + sessionInfo.txt, requirements.txt / environment.yml / uv.lock, Stata version + ado package list), records seeds and RNG kind, optionally writes a pinning…
grant-proposal
Scaffold a research grant proposal (NSF, NIH, ERC, or foundation) by composing existing primitives — pulls identification strategy from an /interview-me spec, delegates the data-management plan to /data-management-plan and the facilities statement to /capture-environment, and emits a funder-requirements checklist. Use…