← 返回 Skills 市场
Pilot Map Reduce
作者
Calin Teodor
· GitHub ↗
· v1.0.0
· MIT-0
79
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install pilot-map-reduce
功能描述
Distributed map-reduce over agent swarms for parallel data processing. Use this skill when: 1. You need to process large datasets across multiple workers 2....
安全使用建议
This skill appears to implement a legitimate map-reduce pattern for the Pilot protocol, but there are gaps and fragile assumptions you should address before installing or running it:
- Verify pilotctl: install pilotctl from a trusted source (homepage or official releases) and inspect its config/credentials. The skill assumes a running pilotctl daemon but gives no guidance on auth or where credentials live.
- Resolve the /tmp inconsistency: the SKILL.md sometimes reads map results from pilotctl output and sometimes from /tmp/map-results-$JOB_ID.json — clarify which approach to use or add the step that writes the file. Running the provided scripts as-is may fail or silently process incomplete data.
- Test in a sandbox: run the workflow on a small, non-sensitive dataset and with a controlled set of worker peers to confirm message formats and persistence behavior.
- Watch for JSON injection/formatting issues: the scripts embed shell variables and jq output into JSON payloads; malformed payloads can break reducers or leak unexpected fields. Quote and validate JSON payloads before sending.
- Check data exposure risk: messages go to an agent swarm; ensure workers are trusted and that payloads do not contain secrets. Confirm how pilotctl persists or exposes received messages (logs, files, sockets).
- Validate dependencies and license: the SKILL.md names jq and sort as dependencies — ensure these are available. Note the AGPL-3.0 license and evaluate whether code snippets or derived work trigger obligations.
- Confirm provenance: the registry metadata owner and source are unknown; verify the pilotprotocol.network homepage and vulture-labs author identity if provenance matters.
If you cannot confirm pilotctl's configuration and the origin of this skill/tooling, treat it carefully and prefer isolated testing rather than deploying on production systems.
功能分析
Type: OpenClaw Skill
Name: pilot-map-reduce
Version: 1.0.0
The skill 'pilot-map-reduce' facilitates distributed data processing via the 'pilotctl' utility (associated with pilotprotocol.network). It is classified as suspicious due to multiple shell-to-jq injection vulnerabilities in SKILL.md, where shell variables such as $JOB_ID and $key are unsafely interpolated directly into jq command strings. While the logic aligns with the stated MapReduce purpose, the combination of insecure command construction and the high-risk nature of agent-to-agent network communication presents a significant vulnerability surface for potential exploitation.
能力评估
Purpose & Capability
Name/description align with the runtime instructions: all commands use pilotctl to discover peers, send messages, and collect results. The declared required binary (pilotctl) is appropriate. Dependencies list jq and sort which are used in the SKILL.md but are not declared as required binaries in the registry metadata (minor mismatch). The homepage and pilot-protocol dependency are consistent with a distributed agent framework.
Instruction Scope
Instructions instruct the agent to list peers, send messages, and read received messages via pilotctl — all within the stated purpose. However there are problematic gaps: the 'Shuffle and reduce' section reads MAP_RESULTS from /tmp/map-results-$JOB_ID.json but earlier collection uses pilotctl --json received (no step writes the /tmp file). The scripts also construct JSON payloads by embedding shell-expanded jq output which is fragile and may produce invalid JSON or accidentally include sensitive content. The SKILL.md assumes pilotctl daemon is running and accessible but doesn't explain authentication, config paths, or how messages are persisted, which affects security and data flow assumptions.
Install Mechanism
Instruction-only skill with no install spec and no code files — lowest-risk in terms of arbitrary code being written to disk. It does, however, require pilotctl and common CLI tools be present on PATH.
Credentials
No environment variables or credentials are requested by the registry metadata. That is reasonable for a CLI-driven skill, but pilotctl may itself rely on local config, sockets, or credentials (not declared here). The skill also references /tmp for temporary results (local filesystem access) and the agent network via pilotctl; you should verify pilotctl's configuration and any keys it uses before running.
Persistence & Privilege
The skill does not request always:true and does not modify other skills or system-wide settings. It runs commands at invocation time only. No elevated persistence behavior is present in the package itself.
如何使用
- 确保已安装 OpenClaw(本地或 Docker 部署)
- 在对话框中输入安装命令:
/install pilot-map-reduce - 安装完成后,直接呼叫该 Skill 的名称或使用
/pilot-map-reduce触发 - 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
Initial release
元数据
常见问题
Pilot Map Reduce 是什么?
Distributed map-reduce over agent swarms for parallel data processing. Use this skill when: 1. You need to process large datasets across multiple workers 2.... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 79 次。
如何安装 Pilot Map Reduce?
在 OpenClaw 或 Claude Code 对话框中运行命令「/install pilot-map-reduce」即可一键安装,无需额外配置。
Pilot Map Reduce 是免费的吗?
是的,Pilot Map Reduce 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。
Pilot Map Reduce 支持哪些平台?
Pilot Map Reduce 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 Pilot Map Reduce?
由 Calin Teodor(@teoslayer)开发并维护,当前版本 v1.0.0。
推荐 Skills