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.
git clone --depth 1 https://github.com/DropFan/claude-code-pluginsWrote 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/plugins/dropfan/claude-code-plugins/lark)<a href="https://agentmods.dev/plugins/dropfan/claude-code-plugins/lark"><img src="https://agentmods.dev/badge/plugins/dropfan/claude-code-plugins/lark/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/plugins/dropfan/claude-code-plugins/lark"><img src="https://agentmods.dev/badge/plugins/dropfan/claude-code-plugins/lark.svg" alt="Reviewed on agentmods" width="80" height="20"></a>Grade A, and why
lark 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.
What it actually says
{
"name": "lark",
"description": "Operate Feishu/Lark (IM, Docs, Base, Sheets, Calendar, Drive, Wiki, Task, OKR, VC, etc.) via the local lark-cli. Mirrors the Claude Code lark skill set for Cowork. 飞书/Lark 全套技能。",
"version": "1.0.96",
"author": {
"name": "Tiger",
"url": "https://github.com/DropFan"
},
"license": "MIT"
}
What it installs
The manifest is a name and a version. 28 skills travel with it, and installing the plugin installs all of them — 3,041 tokens a session between them. Each is measured on its own page, and each can be installed alone.
- Skill lark-openapi-explorer A 76 tokens
- Skill lark-workflow-standup-report A 53 tokens
- Skill lark-skill-maker A 46 tokens
- Skill lark-attendance A 19 tokens
- Skill lark-task A 125 tokens
- Skill lark-im A 140 tokens
- Skill lark-workflow-meeting-summary A 55 tokens
- Skill lark-markdown A 90 tokens
- Skill lark-calendar A 92 tokens
- Skill lark-drive A 208 tokens
- Skill lark-event A 118 tokens
- Skill lark-sheets A 297 tokens
- Skill lark-contact A 120 tokens
- Skill lark-wiki A 169 tokens
- Skill lark-apps A 347 tokens
- Skill lark-base A 131 tokens
- Skill lark-slides A 156 tokens
- Skill lark-okr A 87 tokens
- Skill lark-meeting A 118 tokens
- Skill lark-approval A 91 tokens
- Skill lark-shared A 49 tokens
- Skill lark-whiteboard A 104 tokens
- Skill lark-vc-agent A 37 tokens
- Skill lark-doc A 124 tokens
- Skill lark-note A 33 tokens
- Skill lark-minutes A 35 tokens
- Skill lark-vc A 35 tokens
- Skill lark-mail B 86 tokens
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 Changed 86125885460d
- 1mo ago First seen · 11 lines scan A 667322067bd0
lark is a plugin published in the GitHub repository DropFan/claude-code-plugins (8 stars, last pushed 26d ago), licensed MIT. 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
tlive
tlive hooks + skill + commands — IM approvals, session monitoring, web terminal.
hope
Thinking discipline for AI work. Clarify intent, shape approach, consult experts, guard unsupervised work, hand work back so you own it. The slop hook needs CLAUDECODEENABLEFUNCTIONHOOKS=1.
sound
Code taste as installable rules. Probe a repo's stack, select only the rules its code shows a surface for (evidence-cited per rule), confirm, and scaffold a browsable .claude/sound/ rule map (surfaced by the sound:prime skill). Installed rules are user-owned; re-runs reconcile.
hunch
A grab bag of hunches — experimental skills and plugin features tried here before they earn a place in hope or sound. Everything graduates or dies. The tend mod needs CLAUDECODEENABLEFUNCTIONHOOKS=1.
hold
Comprehension at the human's pace. Grasp a change, drive a behavior, land one thing in an audience — until you can say how it works with the agent out of the room.
director-mode-lite
Guidance-first Director workflows for Claude Code, Codex CLI, and Grok Build, with portable session relay and native adapters.