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 oubakiou/mdxg-redline --skill md-reviewgit clone --depth 1 https://github.com/oubakiou/mdxg-redlineWrote 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/oubakiou/mdxg-redline/md-review)<a href="https://agentmods.dev/skills/oubakiou/mdxg-redline/md-review"><img src="https://agentmods.dev/badge/skills/oubakiou/mdxg-redline/md-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/oubakiou/mdxg-redline/md-review"><img src="https://agentmods.dev/badge/skills/oubakiou/mdxg-redline/md-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 | $0.00178 | $0.03524 |
| Opus 5 | $0.00089 | $0.01762 |
| Sonnet 5 | $0.00036 | $0.00705 |
| Haiku 4.5 | $0.00018 | $0.00352 |
Grade A, and why
md-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 11d 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 — 245 lines — stays where its author put it; the contents beside it link to each section on GitHub.
md-review
A skill that lets Claude run the standard review loop end-to-end using the mdxg-redline CLI. One round consists of the following steps:
- Prepare the target markdown (read an existing .md, have Claude generate one, or use the path specified by the user via
$ARGUMENTS) - Run
scripts/request-review.sh <input.md>to generate the review HTML and launch the browser; capture theREVIEW_HTML/FEEDBACK_JSONpaths from stdout - Run
scripts/wait-feedback.sh <feedback.json>to poll until the feedback.json appears - Read the detected feedback.json and apply the comments to the markdown
- Send the updated markdown out for another review
- Ask the user whether to proceed to the next round, clean up review artifacts and finish, or finish as-is
The reviewer needs time to write comments (minutes to hours), so Claude's turn is free during polling — either work on other tasks or sleep.
Trigger patterns
- Auto-triggering: When the user says things like "send this for review", "I want a human review", "mdxg-redline", "redline", etc., Claude automatically selects this skill based on the description
- Slash command: When the user explicitly runs
/md-review path/to/markdown.md, that argument is used asINPUT_MD. If the argument is empty or natural language, Claude identifies the markdown from context - Natural language path: When the user says something like "send
docs/spec.mdout for review" with an embedded path, that path is used asINPUT_MD
When to use / not use
Use
- When the user explicitly requests structured feedback from a human — e.g. "have someone review this spec", "get comments via mdxg-redline"
- When the user wants to run multiple rounds of review → revision → re-review
- When Claude has generated markdown and the user wants a human check rather than self-review
Do not use
- When the reviewer is an LLM, not a human (this skill processes structured JSON feedback written by human reviewers; it is not for LLM-as-judge style automated evaluation)
- For simple markdown operations that don't involve human review — e.g. "summarize this markdown", "fix the formatting"
What ships with it
7 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.
- evals/evals.json 1.9 KB
- evals/files/eval1-basic/spec-287b2ae21f99425a-feedback.json 618 B
- evals/files/eval1-basic/spec.md 616 B
- evals/files/eval3-mixed/design-43ffe960c0f6a216-feedback.json 1.4 KB
- evals/files/eval3-mixed/design.md 438 B
- scripts/request-review.sh 923 B runs code
- scripts/wait-feedback.sh 265 B 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.
- 11d ago First seen · 245 lines · 178 tokens per session scan A e01a8d086783
md-review is a skill published in the GitHub repository oubakiou/mdxg-redline (13 stars, last pushed 1mo ago), licensed MIT. It adds 178 tokens to every session and 3,524 once invoked, about $0.0009 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 2026-08-30.
Other skills, from other repositories
human-loop-audit
Interaktives Reißverschluss-Audit-Verfahren für eine Serie von Apps/Prüfobjekten: der User testet live und gibt kurzes Feedback im Chat, der Agent startet dabei bereits das nächste Objekt UND wertet parallel das Feedback des vorigen aus (strukturiert erfassen, Reparaturaufträge sofort an Worker delegieren) — statt…
trulens-evaluation-setup
Configure feedback functions and selectors for TruLens evaluations.
firecrawl-build-scrape
Integrate Firecrawl /scrape into product code for single-page extraction. Use when an app already has a URL and needs markdown, HTML, links, screenshots, metadata, or structured page output. Prefer this skill over broader crawl patterns when the feature is page-level.
firecrawl-build-search
Integrate Firecrawl /search into product code and agent workflows. Use when an app needs discovery before extraction, when the feature starts with a query instead of a URL, or when the system should search the web and optionally hydrate result content.
firecrawl-build-interact
Integrate Firecrawl /interact into product code for dynamic pages and browser actions after scraping. Use when a feature needs clicks, form fills, pagination, authentication-aware flows, or other multi-step interactions that plain /scrape cannot complete.
crit-story
Author a crit story and continue the interactive review loop only when the user explicitly invokes crit-story or directly asks you to generate a crit story. Do not infer this skill from generic review, PR, or diff-review requests.