Agent Residue Evidence
Agent Residue Evidence
Agent Residue Evidence (ARE) 为智能体的测试和构建工作流添加了一个证据检查点:
observe → Agent reviews evidence → user authorizes cleanup
→ Agent cleans with its own tools → ARE verifiesARE 显示任务范围内的文件、目录、关联进程及其监听端口。它从不删除文件、停止进程、关闭端口、判定某些内容可安全删除、扫描整台机器、上传数据、使用网络或收集遥测数据。
访问项目站点、运行可复现的任务残留演示,或下载最新已验证版本。
标准任务
在首次测试或构建之前,智能体调用 begin_task_observation,并传入确切的工作区以及任何任务拥有的临时根目录。事件是可选的;可以通过 append_task_events 发送安全的通用事件或空心跳。在给出最终答复之前,智能体调用 end_task_observation,解释候选对象,并询问用户如何处理。只有在获得明确授权后,智能体才会使用其常规工具进行清理。随后它调用 verify_task_residue 重新检查相同的稳定候选对象。
NO_CANDIDATES_OBSERVED 仅表示在声明的观察范围内没有出现候选对象。它并非对整个主机的洁净性声明。
Related MCP server: flameox
无基线的已完成任务
inspect_completed_task 接受显式根目录和有界时间窗口。结果始终为 PARTIAL_EVIDENCE;它可以使用事件、凭证、推断或未归因的证据,但绝不会声称 BASELINE_OBSERVED。使用 get_residue_report 读取已保存的报告,而无需重新观察。
每次观察都通过不透明的属主能力进行隔离。任务 ID、报告 ID、观察 ID 和候选 ID 仅为显示引用,绝不用于授权访问。状态以不透明文件名静态加密存储;确切路径仅会通过经授权的候选解析调用披露。短期执行器能力只能追加明确允许的事件类型和根别名,不能读取、结束、验证或解析任务。
MCP 接口恰好包含八个工具:
begin_task_observationappend_task_eventsend_task_observationinspect_completed_taskget_residue_reportverify_task_residuedelegate_task_executorresolve_residue_candidate
刻意不提供清理、删除、执行、终止或关闭工具。
本地优先打包
一个发布版本包含原生 CLI/stdio MCP 服务器、轻量级 Agent 插件、MCPB、校验和、来源证明和 SPDX SBOM。参见快速入门和安装。运行时操作完全在本地进行且可离线运行。报告默认保留七天,总容量上限为 100 MB;已保留的报告和活动基线均受到保护。卸载时绝不会静默删除报告。
MCP Registry 命名空间为 io.github.fantasyce/agent-residue-evidence。
原生验收通过强制的原生 CI 任务覆盖 macOS 14+ arm64、Linux amd64 和 Windows 11 amd64。仅交叉编译不会被报告为原生验收。
开发
需要 Go 1.26 或 1.27。
bash scripts/check.sh
bash scripts/run_native_acceptance.shThis server cannot be installed
Maintenance
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Related MCP Servers
- AlicenseAqualityBmaintenanceA local-first, model-neutral MCP server for collecting and normalizing change-scoped release evidence. It provides deterministic Git change summaries, evidence collection, and review bundles for agent review.718Apache 2.0
- AlicenseBqualityAmaintenanceEnables coding agents to query, compare, and audit local profiler traces, benchmarks, memory captures, and execution evidence without uploading code or data, using CLI and MCP interfaces.9334MIT
- AlicenseNot gradedqualityAmaintenanceEnables AI agents to query persistent, build-scoped evidence memory for reverse engineering, combining static analysis, runtime captures, and claims with honest uncertainty. Provides read-only access to an evidence graph and MCP prompts for structured investigations.01Apache 2.0
- AlicenseNot gradedqualityBmaintenanceEnables LLM agents to safely reclaim disk space by deleting build artifact directories like node_modules, .venv, and target, with strong guardrails to prevent accidental or malicious deletion.MIT
Related MCP Connectors
Preflight, approve, and prove consequential agent actions with signed evidence and x402 tools.
Watchdog for unattended AI agents: alerts, evidence checks and a verifiable proof per run.
A paid remote MCP for agent memory MCP, built to return verdicts, receipts, usage logs, and audit-re
Latest Blog Posts
- Who's Calling? MCP Hosts Are an Identity Blind Spot (And the Spec Knows It)By Om-Shree-0709 on .mcpAgent IdentityOAuth 2.1
- Your AI Chatbot Just Exposed Your CEO's Salary to an InternBy Om-Shree-0709 on .Agent IdentityMCP SecurityOAuth Delegation
- Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)By Om-Shree-0709 on .Agentic AiPrompt InjectionWebAssembly
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/fantasyce/agent-residue-evidence'
If you have feedback or need assistance with the MCP directory API, please join our Discord server