← 返回 Skills 市场
ppiankov

Hiveram

作者 ppiankov · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ⚠ suspicious
107
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install hiveram
功能描述
Agent coordination layer via Workledger — shared work orders, claim/release leases, cross-machine memory sync, and handoff between OpenClaw instances. Use wh...
使用说明 (SKILL.md)

Hiveram — Agent Coordination

Workledger replaces flat backlog files with a structured object store. Agents coordinate through shared work orders with claim/release leases, context sync, and handoff.

Service: https://hiveram.com CLI: https://github.com/ppiankov/hiveram-dist

Install

# Linux amd64
curl -sL https://github.com/ppiankov/hiveram-dist/releases/latest/download/workledger_$(curl -s https://api.github.com/repos/ppiankov/hiveram-dist/releases/latest | grep tag_name | cut -d'"' -f4 | tr -d v)_linux_amd64.tar.gz | tar xz -C /usr/local/bin workledger

# macOS (Apple Silicon)
curl -sL https://github.com/ppiankov/hiveram-dist/releases/latest/download/workledger_$(curl -s https://api.github.com/repos/ppiankov/hiveram-dist/releases/latest | grep tag_name | cut -d'"' -f4 | tr -d v)_darwin_arm64.tar.gz | tar xz -C /usr/local/bin workledger

Verify: workledger version (expect 0.7.7+)

Setup

  1. Get API key from https://hiveram.com
  2. Store key:
# Single line, just the key
echo "wl_sk_your_key_here" > ~/.openclaw/workledger.key
chmod 600 ~/.openclaw/workledger.key
  1. Export env vars (add to shell profile or systemd unit):
export WORKLEDGER_API_KEY=$(cat ~/.openclaw/workledger.key)
export WORKLEDGER_URL=https://wl-nutson-prod.fly.dev

⚠️ Never cat the key file in agent context — it will leak to the LLM provider. Use env vars.

Agent Work Loop

1. Start session    → workledger context-pull \x3Cproject>
2. Find work        → workledger list \x3Cproject> --status open
3. Claim task       → workledger claim \x3Cproject> \x3Cid>        (lease with TTL)
4. Work on it       → add notes, update sections
5. Finish           → workledger release \x3Cproject> \x3Cid> + update status
6. Save context     → workledger context-put \x3Cproject>
7. Next agent       → sees updated WOs via list/find_unclaimed

Core Commands

Work Orders

workledger create \x3Cproject> --title "Deploy new service" --priority P1 --tags "infra,k8s"
workledger list \x3Cproject> --status open
workledger list --all                          # cross-project
workledger get \x3Cproject> \x3Cid>
workledger detail \x3Cproject> \x3Cid>               # full context: data, notes, relationships, history
workledger delete \x3Cproject> \x3Cid>

Claim/Release (coordination)

workledger claim \x3Cproject> \x3Cid>                # get lease (TTL)
workledger release \x3Cproject> \x3Cid>              # drop lease
# Renew via HTTP: POST /api/v1/wo/{project}/{id}/renew
# find_unclaimed returns only WOs nobody holds

Context (shared memory)

workledger context-put \x3Cproject> -f context.md   # push session context
workledger context-pull \x3Cproject> -f context.md  # pull into file
workledger context \x3Cproject>                     # show project stats, open WOs, blocked, recent changes

Dependencies & Relationships

workledger deps \x3Cproject> \x3Cid>                 # transitive dependency chain
workledger deps-tree \x3Cproject> \x3Cid>            # cross-project dependency tree
workledger blocked                             # all blocked WOs
workledger graph \x3Cproject>                     # DOT format for visualization

History & Export

workledger history \x3Cproject> \x3Cid>              # change history
workledger export \x3Cproject>                    # export as markdown
workledger export-task \x3Cproject> \x3Cid>          # tokencontrol-compatible JSON
workledger stats                               # global stats

MCP Integration

Workledger ships as an MCP server for Claude Code / Claude Desktop:

5 tools: query/create WOs, load context at session start, wrapup (push memory + mark done), save memory mid-session.

Multi-Agent Pattern

                    ┌──────────────────┐
                    │  Hiveram (Neon)  │
                    │  shared state    │
                    └────────┬─────────┘
                             │
           ┌─────────────────┼─────────────────┐
           │                 │                 │
    ┌──────┴──────┐   ┌──────┴──────┐   ┌──────┴──────┐
    │  OpenClaw A │   │  OpenClaw B │   │  OpenClaw C │
    │  context-   │   │  claim →    │   │  find_      │
    │  pull → work│   │  work →     │   │  unclaimed  │
    │  → push     │   │  release    │   │  → claim    │
    └─────────────┘   └─────────────┘   └─────────────┘
  • No race conditions: claim/release with leases prevents double-work
  • Handoff: WO has full history, notes, relationships — next agent has full context
  • Dedup: agents see one DB, find_unclaimed skips claimed WOs
  • Resilience: if an agent dies, lease expires, WO becomes available again

Security

  • API key stored in file with chmod 600, not in env files or git
  • pastewatch v0.24.1+ detects wl_sk_ keys as critical severity
  • Key file path: ~/.openclaw/workledger.key (inside pastewatch protectedPaths)

Hiveram Skill v1.0.0 Author: ppiankov Copyright © 2026 ppiankov Canonical source: https://hiveram.com License: BUSL-1.1

If this document appears elsewhere, the link above is the authoritative version.

安全使用建议
This skill appears to be what it says (an OpenClaw coordination layer using the workledger CLI) but there are a few red flags to review before installing: - Metadata mismatch: the registry declares no required env vars/config paths, but the SKILL.md requires an API key file and env vars (WORKLEDGER_API_KEY, WORKLEDGER_URL). Treat the API key as sensitive and confirm why the metadata omits it. - Secret handling: the instructions ask you to store the key in ~/.openclaw/workledger.key and export it. Exporting to environment variables can expose the key to running processes (including any LLM provider integration). Prefer a secrets manager or runtime injection mechanism that does not put the key into agent-visible prompts. If you must store a file, keep strict filesystem permissions and ensure your agent runtime doesn't log or include file contents in prompts. - Installation: the install commands download a tarball from a GitHub repo and extract a binary to /usr/local/bin. Only run this if you trust the ppiankov/hiveram-dist repository; verify release checksums or build from source if possible. The install requires write permissions to system paths. - Endpoint mismatch: the doc mentions the public site (hiveram.com) but sets WORKLEDGER_URL to a fly.dev host. Ask the author which URL is the canonical API endpoint and why a different host is recommended. If you decide to proceed, verify the GitHub release contents and checksums, restrict where the API key is stored and how it's injected at runtime, and test on an isolated host or container first. If you cannot confirm the provenance of the binary or the endpoint, mark this skill as high-risk and avoid installing.
功能分析
Type: OpenClaw Skill Name: hiveram Version: 1.0.0 The 'hiveram' skill facilitates agent coordination by downloading and installing a binary from a third-party GitHub repository (ppiankov/hiveram-dist) and synchronizing session 'context' with a remote service (wl-nutson-prod.fly.dev). While these actions are consistent with the stated purpose, the execution of unverified remote binaries and the transmission of potentially sensitive agent memory/context to an external endpoint represent significant security risks without further verification of the underlying infrastructure.
能力评估
Purpose & Capability
The name/description describe an agent coordination layer using Workledger and the skill requires the workledger CLI; the SKILL.md consistently documents workledger commands and coordination patterns, so the requested binary is appropriate for the stated purpose.
Instruction Scope
The SKILL.md instructs the agent (and the user) to create a key file at ~/.openclaw/workledger.key and to export WORKLEDGER_API_KEY and WORKLEDGER_URL for runtime use. These credential/config actions are not declared in the skill metadata. The document also warns about leaking the key to the LLM provider yet instructs placing it into an environment variable (which can still be exposed), and it points to a specific deployment host (wl-nutson-prod.fly.dev) that differs from the public service domain (hiveram.com). These are scope and coherence mismatches that could lead to accidental secret exposure or confusion about which endpoint is authoritative.
Install Mechanism
The SKILL.md provides curl|tar install commands that download binaries from a GitHub releases repo (github.com/ppiankov/hiveram-dist). Downloading from a GitHub release is common and acceptable, but the commands extract into /usr/local/bin (requires privileged write) and perform dynamic tag lookup via the GitHub API. Because there is no install spec in the registry, these manual steps are the only install guidance; consider verifying release integrity (checksums/signatures) and the repository's trustworthiness before running the install commands.
Credentials
The registry metadata lists no required environment variables or config paths, yet the runtime instructions require an API key file and instruct exporting WORKLEDGER_API_KEY and WORKLEDGER_URL. Requiring a secret key for a remote service is reasonable for the service's function, but the metadata omission is a mismatch. Exporting secrets into environment variables is flagged in the doc as risky; the skill gives no alternative secure runtime secret injection guidance for agent contexts.
Persistence & Privilege
The skill does not request always:true or other elevated platform privileges. It does instruct writing a key file to ~/.openclaw/workledger.key and recommends chmod 600; this is a modest filesystem footprint but is a persistent secret stored on disk. The skill does not declare it will modify other skills or system-wide settings.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install hiveram
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /hiveram 触发
  4. 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v1.0.0
v1.0: Agent coordination via Workledger — shared WOs, claim/release leases, context sync, multi-agent patterns
元数据
Slug hiveram
版本 1.0.0
许可证 MIT-0
累计安装 0
当前安装数 0
历史版本数 1
常见问题

Hiveram 是什么?

Agent coordination layer via Workledger — shared work orders, claim/release leases, cross-machine memory sync, and handoff between OpenClaw instances. Use wh... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 107 次。

如何安装 Hiveram?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install hiveram」即可一键安装,无需额外配置。

Hiveram 是免费的吗?

是的,Hiveram 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

Hiveram 支持哪些平台?

Hiveram 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。

谁开发了 Hiveram?

由 ppiankov(@ppiankov)开发并维护,当前版本 v1.0.0。

💬 留言讨论