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 unbias38/my-claude-skills --skill paper-language-passgit clone --depth 1 https://github.com/unbias38/my-claude-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/unbias38/my-claude-skills/paper-language-pass)<a href="https://agentmods.dev/skills/unbias38/my-claude-skills/paper-language-pass"><img src="https://agentmods.dev/badge/skills/unbias38/my-claude-skills/paper-language-pass/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/unbias38/my-claude-skills/paper-language-pass"><img src="https://agentmods.dev/badge/skills/unbias38/my-claude-skills/paper-language-pass.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 | $0.00276 | $0.06719 |
| Opus 5 | $0.00138 | $0.03359 |
| Sonnet 5 | $0.00055 | $0.01344 |
| Haiku 4.5 | $0.00028 | $0.00672 |
Grade A, and why
paper-language-pass 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 — 329 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Paper Language Pass
When to use
The user has a complete academic manuscript whose science is already validated (peer review done, reviewers' substantive concerns addressed) and wants a language-layer polish across the entire paper. Eight specialist subagents run in parallel, each scanning the whole paper for a single dimension of writing quality:
| Pass | Subagent | Looks for |
|---|---|---|
| 1 | Consistency auditor | Term/acronym/capitalization/number/unit/hyphenation consistency |
| 2 | Tense auditor | Section-appropriate tense, within-section consistency, et al. agreement |
| 3 | Hedging auditor | Claim-strength calibration, overclaiming, under-hedging, abstract-body alignment |
| 4 | Prose polisher (review only) | Grammar, nominalization, fillers, wordiness, awkward phrasing — sentence level only |
| 5 | Coherence reviewer | Paragraph claim-first, transitions, argument chaining, organizational redundancy |
| 6 | Abstract auditor | WHY→PROBLEM→HOW→RESULTS structure, acronyms, no citations, single paragraph |
| 7 | Manuscript hygiene auditor | Three classes of phrasing that leak from adjacent genres into the manuscript body and never belong there: (A) reviewer / editor process meta-discourse, (B) journal self-reference / sycophancy, (C) implementation / engineering-detail leakage including untaken fallback paths |
| 8 | AI-authorship tell auditor | Eight tell families measured as whole-manuscript density: hype register (self-praise + literature disparagement), self-coined theoretical jargon, evaluative adverb-comma openers, procedural section roadmaps, near-verbatim argument recycling, em-dash density, signature vocabulary, mechanical rhetorical templates. Reports a signature score (families fired out of 8) |
Why AI tells get their own pass. These patterns are only meaningful as density and co-occurrence measurements across the whole manuscript. One Notably, is a word; eleven is a habit. One em dash is punctuation; 4.6 per thousand words is a fingerprint. A sentence-level reviewer cannot tell which one it is looking at, so Passes 1–7 have been explicitly instructed to leave every AI tell to Pass 8, and Pass 8 has been given per-family firing thresholds so it does not blanket-flag normal English. If the user's concern is specifically "does this read as machine-written", Pass 8 is the pass that answers it.
What ships with it
9 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.
- README.md 8.4 KB
- references/pass1-consistency.md 4.6 KB
- references/pass2-tense.md 5.1 KB
- references/pass3-hedging.md 7.5 KB
- references/pass4-prose.md 6.4 KB
- references/pass5-coherence.md 12 KB
- references/pass6-abstract.md 6.0 KB
- references/pass7-manuscript-hygiene.md 13 KB
- references/pass8-ai-authorship-tells.md 31 KB
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 · 329 lines · 276 tokens per session scan A e9c787a4908c
paper-language-pass is a skill published in the GitHub repository unbias38/my-claude-skills (2 stars, last pushed 17d ago), licensed MIT. It adds 276 tokens to every session and 6,719 once invoked, about $0.0014 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
data-warehouse-experimentation
Running experiments out of the data warehouse instead of via dedicated experiment platforms. SQL-based assignment, exposure logging discipline, metric definitions in dbt models, statistical analysis in SQL or Python, variance reduction with CUPED, sequential testing, and the operational tradeoffs vs platforms like…
data-scientist
Processes and analyzes data with resident-kernel engines (DuckDB, Polars) and one-shot tools. Use for CSV/parquet/JSON analysis, group-by/join/aggregation, time series, distributions, cleaning, or plotting a dataset.
thinking-scientific-method
When a symptom has several plausible causes, rank falsifiable hypotheses and run the cheapest discriminating observation first; prefer least-assumptive survivors only after evidence fit.
crucible-research-foundations
Validate findings, design shuffled nulls, check label leakage, review causal features. TRIGGERS - shuffled null, label leakage.
arxiv-source-first-paper-ingest
Read an arXiv paper faithfully by fetching the authors' own LaTeX source instead of OCR-ing the PDF. Extracts every display equation with its label, environment and enclosing proposition, renders each to MathML, and proves usability by compiling each formula with a real TeX engine. Use when you need a paper's formulas…
hypothesis-testing-engine
Take any claim and design + execute a complete research protocol to test it. Apply scientific method automatically: design study, gather data, run analysis, provide verdict with confidence level.