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 emaballarin/ccplugins --skill paper-reviewgit clone --depth 1 https://github.com/emaballarin/ccpluginsWrote 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/emaballarin/ccplugins/paper-review)<a href="https://agentmods.dev/skills/emaballarin/ccplugins/paper-review"><img src="https://agentmods.dev/badge/skills/emaballarin/ccplugins/paper-review/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/emaballarin/ccplugins/paper-review"><img src="https://agentmods.dev/badge/skills/emaballarin/ccplugins/paper-review.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 | $0.00092 | $0.02626 |
| Opus 5.5 | $0.00037 | $0.01050 |
| Sonnet 5.5 | $0.00018 | $0.00525 |
| Haiku 4.5 | $0.00009 | $0.00263 |
Grade A, and why
paper-review 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 25d 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 — 94 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/ccsci:paper-review — second reader for a human-written review
The review is the reviewer's. What Claude adds is the one thing a sequential human reader does not have: the whole paper in view at once, so a claim in the abstract can be held against the table meant to support it, a disclaimer on one line against the hypothesis three lines above it, an announced four-way comparison against the numbers actually reported. The chief tools are recap tables, schematic decompositions, reminders of what is missing, and error checks on the reviewer's drafts. Prose for the review form comes from Claude only when the reviewer says "draft it".
Everything stays in the conversation. Files are written on request only (§5).
Pacing. The phases are turns, not one message. Stop after §0 for the PDF and the impression; after §3 for the score, because the block table of §4 depends on the band; and after each deliverable's constraints for the reviewer's draft. Run end to end, §0–§4 is five thousand words — right, and unreadable; paced, each turn is one thing the reviewer can act on.
0. Frame
Fix the frame before opening the PDF, because it decides the shape of every deliverable:
| Field | Why it matters |
|---|---|
| Venue and track | Norms differ: a four-page workshop paper is not judged as a main-track submission. |
| Rubric | The score scale with its boundary definitions. Ask the reviewer to paste the official text; when only the venue is named, search for its current reviewer guidelines and quote what is found, with the source. A rubric that cannot be sourced is labelled unsourced and stays labelled at every calibration step — an invented rubric calibrates nothing. |
| Rebuttal | No rebuttal ⇒ a declarative review: every question becomes a statement of what the paper does not say. |
| Page limit and appendix | Decides whether appendix-only evidence counts toward a main-text claim. |
| Form fields | Which boxes exist — summary, one-line title, strengths/weaknesses, questions, confidence, confidential comment — so nothing is written for a box that is not there. |
| Siblings | Other papers in the same batch, for calibration on request (§3). |
What ships with it
3 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.
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.
- 25d ago First seen · 94 lines · 92 tokens per session scan A 7677519fe02d
paper-review is a skill published in the GitHub repository emaballarin/ccplugins (4 stars, last pushed 10d ago), licensed MIT. It adds 92 tokens to every session and 2,626 once invoked, about $0.0004 per session on Opus 5.5. 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 2026-09-24.
Other skills, from other repositories
r-mirai
Help users write correct R code for async, parallel, and distributed computing using mirai. Use when users need to run R code asynchronously or in parallel, write mirai code with correct dependency passing, set up parallel workers, convert from future or parallel, use miraimap, integrate with Shiny or promises, or…
r-cran-status
Look up an R package's live status on cran.r-project.org - submission/review state (queue, human review, waiting, archived, past version's fate) or R CMD check results (OK/NOTE/WARN/ERROR per platform). Use for "what's the CRAN status of X", "did X get accepted/rejected", "is X passing CRAN checks", "when was version…
paper-opportunity-radar
Run a cumulative daily or retrospective sweep of research papers on a chosen topic, audit their claims, methods, integrity signals, and independent support, then identify overlooked but feasible project or business opportunities in a detailed source-grounded report. Use when asked to monitor papers every day, mine…
aql-authoring
This skill should be used when the user asks to "write an AQL query", "optimize an AQL query", "review AQL", or "query openEHR data" — the multi-step authoring/optimization workflow for AQL (Archetype Query Language) over openEHR clinical data. For a one-off explanation of an existing query or a single AQL…
ant-colony-optimization
When the user wants to design, implement, or tune ant colony optimization for combinatorial problems, including pheromone model design, visibility heuristics, and choosing among Ant System, Ant Colony System, and MAX-MIN Ant System. Also use when the user mentions "ant colony," "ACO," "pheromone," "MMAS," "ant colony…
benders-decomposition
When the user wants to solve a structured MILP or two-stage stochastic program by Benders decomposition — splitting it into an integer master and LP subproblems, deriving optimality and feasibility cuts from subproblem duals, and implementing either the classic iterative loop or branch-and-Benders-cut with…