agent-skills
mrshu/agent-skills
Useful skills for AI agents (Claude Code, etc.)
Instalação
npx convez-ai mrshu/agent-skills1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.
Métricas
README
A tradução para o português deste README ainda está na fila.
agent-skills
A Claude Code plugin marketplace with useful skills for AI agents.
Installation
First, add the marketplace:
/plugin marketplace add mrshu/agent-skills
Then install plugins:
/plugin install scholar-search
/plugin install wshot
/plugin install paper-siphon
/plugin install codex-exec
/plugin install claude-exec
/plugin install review-anvil
/plugin install overleaf-comment
/plugin install overleaf-compile
/plugin install gdocs-comment
Plugins
scholar-search
Find academic papers, explore citations, and export BibTeX. Uses four tools under the hood:
- s2cli (Semantic Scholar) — paper recommendations, citation/reference networks, arXiv/DOI lookups
- openalexcli (OpenAlex) — broadest coverage (260M+ works), institution/journal queries, aggregation
- arxivy (arXiv) — direct arXiv search, paper lookup by ID, browse recent submissions by category
- dblpcli (DBLP) — high-quality, curated BibTeX export
All tools run via uvx — no installation needed.
paper-siphon
Convert academic PDFs to clean Markdown. Supports local files and remote URLs (e.g. arXiv). Three extraction modes:
- Default — fast Docling-based extraction, handles most papers
- VLM (
--vlm) — vision-language model for complex layouts and scans - Formula enrichment (
--enrich-formula) — better LaTeX fidelity for math-heavy papers
Runs via uvx — no installation needed.
codex-exec
Delegate code review, plan review, and deep exploration tasks to Codex CLI from Claude Code. Use it as a pre-push quality gate, plan counter-review, or deep-dig exploration tool. Key commands:
codex review --base main— review current branch against maincodex review --uncommitted— review work-in-progress changescodex exec '<prompt>'— custom prompted review, plan critique, or codebase exploration
claude-exec
Delegate code review, plan review, and deep exploration tasks to Claude Code CLI from Codex or other agents. The mirror of codex-exec — lets non-Claude agents get an independent second opinion. Key commands:
claude -p '<prompt>'— non-interactive review or explorationclaude -p --permission-mode plan '<prompt>'— read-only safe review mode
review-anvil
Iterative multi-agent review-and-fix loop. Wraps the "let's do three rounds of fix/review" pattern: dispatch parallel codex-exec and claude-exec reviewers, synthesize their findings, apply fixes with logically-separated commits, and repeat. Defaults to 3 rounds × 3 reviewers (2 codex + 1 claude). Configurable rounds, agent count/mix, focus, and target.
The plugin ships four cross-agent skills under plugins/review-anvil/skills/:
review-anvil— the engine. The actual review loop with the full parameter surface (rounds,agents,focus,target,min_fix_severity,allow_new_deps,commit_mode,report_path). The default mode is fix-and-commit (commit_mode=per_fix).review-anvil-readonly— preset. Read-only review pass: activates the engine withcommit_mode=noneand a default ofrounds=1. No edits, no commits.review-anvil-pr— preset. Read-only review of a GitHub PR (github.com or GitHub Enterprise) with hybrid posting: findings with afile+lineanchor become inline review comments threaded next to the code; findings without anchors (architectural / overview) roll into a top-level summary body. The whole thing submits as one PR review (one timeline event). Falls back to a top-level comment if no findings are anchored or if GitHub rejects the inline payload (e.g. reviewer-supplied line numbers outside the diff). Pairs the engine in read-only mode withscripts/pr-helper.shfor locator parsing,ghpreflight, and posting. Auto-detects the PR from the current branch when no locator is passed. Requiresghandjqon PATH.review-anvil-improve-pr— preset. Productive PR loop with a "narrate then update" UX: (1) post a top-level "starting" PR comment cc'ing the original author (notifying them what's about to happen), (2) review + apply fix commits across N rounds, (3)git pushback to the PR, (4) PATCH-edit the starting comment in-place with the synthesized report (success) or a failure summary (any failure). One comment in the PR timeline, two states, one notification. Reusespr-helper.shwithverify-checkout/post-start/post-updatesubcommands. Targets the local branch directly (<base>...HEAD) to bypass the engine's "PR-target / per_fix incompatibility" rule. Auto-detects PR from current branch when no locator is passed. Requiresghon PATH; you must be on the PR's branch with a clean worktree.
Four skills (not four slash commands) because skills are the cross-agent abstraction — anything that supports Skill <name> (Claude Code, Codex CLI, Cursor, OpenCode, Continue, Cline, Gemini CLI, …) can activate any of these by description match. The npx skills add mrshu/agent-skills --all install carries all four SKILL.mds and the shared helper script under review-anvil-pr/scripts/. Multi-forge support (GitLab MR, Gitea PR, …) is a v2 concern.
In Claude Code, you can also invoke directly as Skill review-anvil "<free-form args>" — see skills/review-anvil/SKILL.md for the full parameter surface.
overleaf-comment
Bulk-post Overleaf comments from a JSON plan. Drives the user's already-logged-in browser tab via Chrome DevTools Protocol, so no API keys / no test accounts. Break-even is around 5 comments — under that, type them by hand.
overleaf-comment --list— show open Overleaf project tabsoverleaf-comment <target> plan.json— post the plan against a tab prefix from--listoverleaf-comment --url https://www.overleaf.com/project/<id> plan.json— auto-resolve the tab from the project URL
Plan items are {file, line, text}. Failed/skipped items are written to <plan>.residual.json so a retry is one command.
Requirements:
- A
chrome-cdpskill installation (auto-discovered or via$CDP_BIN) - Brave / Chrome / Chromium running with
--remote-debugging-port=9222 - The Overleaf project tab open and logged in, with commenting permission
overleaf-compile
Recompile an Overleaf project and extract its build logs/errors/warnings in a standardized format — closing a push → compile → debug loop. API-first: calls Overleaf's internal compile endpoint from inside the logged-in tab and fetches the raw output.log / output.blg, so the diagnostics are complete (not the truncated logs pane).
overleaf-compile --list— show open Overleaf project tabsoverleaf-compile <target>— recompile and emit standardized JSON logsoverleaf-compile --url https://www.overleaf.com/project/<id>— auto-resolve the taboverleaf-compile <target> --errors-only --format compiler— fast human-readable triage
Entries are {level, file, line, message, source, count} with level ∈ error|warning|typesetting|info|bib-*. Note that Overleaf's status: "success" only means a PDF was produced — the exit code (and parsed errors count) is the real verdict. --clear-cache recompiles from scratch; --out-dir saves the raw artifacts; --no-recompile reads the on-screen logs pane without compiling.
Requirements:
- A
chrome-cdpskill installation (auto-discovered or via$CDP_BIN) - Brave / Chrome / Chromium running with
--remote-debugging-port=9222 - The Overleaf project tab open and logged in
python3on PATH (log parsing)
gdocs-comment
Bulk-post anchored Google Docs comments (with @-mentions) from a JSON plan. Drives the user's already-logged-in browser tab via Chrome DevTools Protocol — no Drive API keys, no OAuth dance. Break-even is around 5 comments — under that, type them by hand.
gdocs-comment --list— show open Google Docs tabsgdocs-comment <target> plan.json— post the plan against a tab prefix from--listgdocs-comment --url https://docs.google.com/document/d/<id>/edit plan.json— auto-resolve the tab from the document URLgdocs-comment --dry-run plan.json— validate the plan without posting
Plan items are {anchor, text, mentions?} — anchor is a substring of the doc body, and the comment is anchored to the first match. Failed items are written to <plan>.residual.json for one-command retry.
Requirements:
- A
chrome-cdpskill installation (auto-discovered or via$CDP_BIN) - Chrome / Brave / Edge with remote debugging enabled
- The Google Doc tab open and logged in, with comment permission
- Node.js 22+
wshot
Screenshot tool for Wayland/GNOME. Capture windows by app name, PID, title, or window ID.
Requirements:
- GNOME with Wayland
window-callsextension (auto-prompted if missing)
License
MIT