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.
/plugin marketplace add MidnightDarling/collatenpx agentmods add plugins/midnightdarling/collate/marketplacegit clone --depth 1 https://github.com/MidnightDarling/collateGrade A, and why
collate 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 2d 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.
What it actually says
{
"name": "collate",
"owner": {
"name": "Alice",
"email": "[email protected]"
},
"metadata": {
"description": "点校 · Collate — Read, think, and write history with your agents. 历史文本 OCR 校对 + 公众号排版 + 学术场研究的 agent 工作流。"
},
"plugins": [
{
"name": "collate",
"source": "./",
"description": "历史论文 OCR 校对工作流：PDF 预处理去水印去角标、MinerU OCR、三类文献（繁体古籍 / 民国排印 / 现代简体）校对、公众号排版导出。支持 Claude Code 与 Codex。",
"version": "0.2.0",
"author": {
"name": "Alice",
"email": "[email protected]"
},
"homepage": "https://github.com/MidnightDarling/collate",
"repository": "https://github.com/MidnightDarling/collate",
"license": "Apache-2.0 AND CC-BY-4.0",
"keywords": [
"ocr",
"chinese-history",
"proofreading",
"mineru",
"wechat-mp",
"traditional-chinese",
"republican-era",
"academic-workflow",
"cross-runtime"
],
"category": "workflow",
"tags": [
"ocr",
"history",
"academic",
"proofreading",
"chinese"
],
"strict": false
}
]
}
What ships with it
1 file beside marketplace.json 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.
- 2d ago First seen · 46 lines scan A ea07dda30936
collate is a plugin published in the GitHub repository MidnightDarling/collate (6 stars, last pushed 3mo ago), licensed Apache-2.0. Its token cost is not measured: this kind of file is read by the harness, not the model. 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 plugins, from other repositories
ctx
Plugin marketplace listing 1 plugin: ctx.
ctx
Fast local search and 'git blame' for agent sessions.
honey-for-devs
Honey (I Shrunk the AI) — three reflexive levers that cut agent token cost: less code (YAGNI/stdlib-first), less prose, and denser agent-to-agent handoffs (ESON/compact-JSON). Plus on-demand satellites (review, eco, gain, compress) and a hive of read-only subagents that return compressed handoffs. Auto-intensity…
ai-dev-kit
DEPRECATED — the databricks-ai-dev-kit plugin is no longer published. Install skills via install.sh or databricks aitools install, which installs the official databricks plugin (Claude Code/Codex/Copilot) or raw skill files (Cursor/OpenCode/Antigravity). Not a move away from plugins. See .claude-plugin/DEPRECATED.m.
powabase-agent-skills
Plugin marketplace listing 1 plugin: powabase.
powabase
Build on Powabase — the AI Backend-as-a-Service (RAG, agents, orchestration, workflows) plus a Supabase-style BaaS layer. Bundles the powabase agent skill under skills/.