← 返回 Skills 市场

agent-exporter Archive Governance Workbench

Name: agent-exporter Archive Governance Workbench
Author: xiaojiou176

作者 Yifeng[Terry] Yu · GitHub ↗ · v1.0.0 · MIT-0

cross-platform ✓ 安全检测通过

116

总下载

当前安装

版本数

在 OpenClaw 中安装

/install agent-exporter-archive-governance-workbench

功能描述

Use when an agent needs to wire the local agent-exporter MCP bridge, publish a local archive shell, save retrieval reports, or read governance evidence and p...

使用说明 (SKILL.md)

agent-exporter Archive Governance Workbench

Use this skill when an agent needs to attach the local agent-exporter bridge from a repo checkout and operate the archive/retrieval/governance workflow through a host-native packet.

What this skill teaches

how to attach the local stdio MCP bridge
how to prove the bridge on a safe first-success path
how to publish a local archive shell
how to save local retrieval reports
how to read governance evidence without pretending there is a hosted platform

Product truth

agent-exporter is a local-first archive and governance workbench
the current primary surface is still CLI-first
this public skill packet is a secondary public lane
the bridge exposes read-mostly archive, retrieval, and governance tools
the browser pages organize proof; they do not execute retrieval or onboarding

First-success flow

Follow references/INSTALL.md and wire the bridge from a repo checkout.
Call integration_evidence_policy_list to prove the bridge can answer a no-risk governance read.
If you have a workspace with transcript HTML receipts, call publish_archive_index on that workspace.
Only after the bridge works, move to retrieval or evidence comparison.

Preferred evidence order

references/INSTALL.md
references/DEMO.md
references/CAPABILITIES.md
references/TROUBLESHOOTING.md

Example prompts

"Wire the local agent-exporter bridge and show me the available governance policies."
"Publish the archive shell for this workspace and tell me where the archive shell landed."
"Run semantic retrieval for this workspace and save a local report."
"Compare two saved integration evidence snapshots and explain the readiness delta."

Truth language

Good: "local stdio bridge"
Good: "secondary public lane"
Good: "archive and governance workbench"
Forbidden: "hosted platform"
Forbidden: "listed-live skill" without fresh host read-back
Forbidden: "full CLI exposed through MCP"

This packet instructs an agent to run a repo-local Python MCP bridge script and to read and produce local archive/retrieval/governance reports. Before enabling it, verify you trust the repo and the script at /absolute/path/to/agent-exporter/scripts/agent_exporter_mcp.py (or the equivalent in your checkout), since the agent will start that local process. Expect local filesystem access to the repo and written local reports; the skill is read-focused but will execute repo-local binaries (python/cargo). If you are unsure about the repo's safety, inspect the referenced scripts and the agent-exporter repo (releases/README) first and run in an isolated environment.

功能分析

Type: OpenClaw Skill Name: agent-exporter-archive-governance-workbench Version: 1.0.0 The skill bundle provides instructions and configuration for an AI agent to interface with a local tool called 'agent-exporter' via the Model Context Protocol (MCP). The content consists of documentation (SKILL.md, README.md) and configuration snippets (OPENCLAW_MCP_CONFIG.json) designed to facilitate local archive management and governance reporting. There is no evidence of malicious intent, data exfiltration, or harmful prompt injection; the instructions explicitly emphasize a 'local-first' approach and set clear boundaries against claiming hosted status.

能力标签

crypto

能力评估

✓ Purpose & Capability

Name/description match the requested actions: wiring a local agent-exporter stdio MCP bridge, publishing local archive shells, running retrieval, and reading governance evidence from a repo checkout. Nothing requested (no env vars, no installs) is out of scope for this purpose.

✓ Instruction Scope

SKILL.md restricts work to a repo checkout and local stdio bridge and explicitly prefers a safe first-success read-only governance call (integration_evidence_policy_list). It does reference running a repo-local Python MCP script and accessing workspace files, which is expected for this skill and stays within the stated boundaries (read-mostly).

✓ Install Mechanism

No install spec or remote downloads — instruction-only. The packet provides config snippets and expects local binaries/scripts already present in the repo checkout, which is reasonable for a local-first workbench.

✓ Credentials

No environment variables, credentials, or external config paths are required. The only required access is to a local repo checkout and local python/cargo binaries, which is proportional to the described task.

✓ Persistence & Privilege

always is false and there is no indication the skill attempts to persist or modify other skills or system-wide settings. Autonomous invocation is allowed by default but nothing in the packet leverages elevated persistent privileges.

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 agent-exporter-archive-governance-workbench
安装完成后，直接呼叫该 Skill 的名称或使用 /agent-exporter-archive-governance-workbench 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

agent-exporter-archive-governance-workbench 1.0.0 - Initial release introducing a skill packet for local archive, retrieval, and governance workflows using the agent-exporter MCP bridge. - Enables agents to wire the local stdio bridge, publish archive shells, save semantic retrieval reports, and read governance evidence and policy packs from repository checkouts. - Focuses on local-first, CLI-oriented use; provides guidance for first-success flows and evidence orders. - Clarifies truth language and prohibited claims, ensuring proper differentiation between local and hosted environments. - Includes example prompts and directs users to key reference documentation.

元数据

Slug agent-exporter-archive-governance-workbench

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