Token导航 LogoToken导航TokenDH.com
研究检索敏感数据clawhub未标认证来源可访问clear审计提醒

pastewatch-mcppastewatch MCP 搜索

Agent Skill

pastewatch-mcp 用于查找、检索和筛选相关信息,适合在 OpenClaw 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

总安装

15,758

周安装

670

GitHub Stars

公开资料未说明

下载量

5,521
OpenClaw

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

MIT-0

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:pastewatch-mcp(pastewatch MCP 搜索)
来源仓库:https://github.com/ppiankov/pastewatch-mcp
安装命令:
openclaw skills install pastewatch-mcp
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 OpenClaw 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

ClawHubOpenClaw
openclaw skills install pastewatch-mcp

简介

OpenClaw 代理的秘密检测 MCP 服务器,防止敏感信息泄露。

  • 适用于代码审查和开发环境安全检查。pastewatch-mcp 属于研究检索类 Skill,可作为该场景下的辅助能力补充。
  • 支持识别 API 密钥、SSH 密钥等多种秘密类型。
  • 需配置白名单和排除规则以减少误报。
  • 注意权限控制,避免过度扫描影响性能。

SKILL.md

name
pastewatch-mcp
description
Secret redaction MCP server for OpenClaw agents. Prevents API keys, DB credentials, SSH keys, emails, IPs, JWTs, and 30+ other secret types from leaking to LLM providers. Includes guard command, API proxy, canary tokens, encrypted vault, git history scanning, org posture scanning, file watcher, and dashboard. Use when reading/writing files that may contain secrets, setting up agent security, or auditing for credential exposure.
metadata
{"openclaw":{"requires":{"bins":["pastewatch-cli","mcporter"]}}}

Pastewatch MCP — Secret Redaction

Prevents secrets from reaching your LLM provider. The agent works with placeholders, secrets stay local.

Source: https://github.com/ppiankov/pastewatch

Install

# macOS
brew install ppiankov/tap/pastewatch

# Linux (binary + checksum)
curl -fsSL https://github.com/ppiankov/pastewatch/releases/latest/download/pastewatch-cli-linux-amd64 \
  -o /usr/local/bin/pastewatch-cli
curl -fsSL https://github.com/ppiankov/pastewatch/releases/latest/download/pastewatch-cli-linux-amd64.sha256 \
  -o /tmp/pastewatch-cli.sha256
cd /usr/local/bin && sha256sum -c /tmp/pastewatch-cli.sha256
chmod +x /usr/local/bin/pastewatch-cli

Verify: pastewatch-cli version (expect 0.24.1+)

MCP Server Setup

mcporter config add pastewatch --command "pastewatch-cli mcp --audit-log /var/log/pastewatch-audit.log"
mcporter list pastewatch --schema  # 6 tools

Severity threshold: pastewatch-cli mcp --min-severity medium or set mcpMinSeverity in .pastewatch.json.

Agent Integration (one-command setup)

pastewatch-cli setup claude-code    # hooks + MCP + CLAUDE.md snippet + doctor check
pastewatch-cli setup cline          # MCP + hook instructions
pastewatch-cli setup cursor         # MCP + advisory

--severity aligns thresholds. --project for project-level config. Idempotent — safe to re-run.

MCP Tools

ToolPurpose
pastewatch_read_fileRead file with secrets replaced by placeholders
pastewatch_write_fileWrite file, resolving placeholders back locally
pastewatch_check_outputVerify text contains no raw secrets
pastewatch_scanScan text for sensitive data
pastewatch_scan_fileScan a file
pastewatch_scan_dirScan directory recursively

Placeholder format: __PW_TYPE_N__ (changed from __PW{TYPE_N}__ in v0.19.7 for LLM proxy compatibility). Configurable via placeholderPrefix.

API Proxy — Last Line of Defense

Catches secrets from sub-agents that bypass hooks and MCP. Sits between your agent and the LLM provider, scanning all outbound requests.

v0.24.0+: Alert injection — when secrets are redacted, a [PASTEWATCH] alert is prepended to the API response so the agent gets immediate feedback. Disable with --no-alert.

pastewatch-cli proxy                    # scans all outbound API requests
pastewatch-cli proxy --port 9998 --upstream https://api.anthropic.com
pastewatch-cli proxy --port 9998 --upstream https://api.anthropic.com --no-alert  # silent redaction
pastewatch-cli proxy --forward-proxy http://corporate:8080  # chain with corporate proxy
pastewatch-cli proxy --severity medium --audit-log /var/log/pastewatch-proxy.log

Detected Secret Types (36)

Includes: AWS, Anthropic, OpenAI, Groq, Perplexity, Hugging Face, Workledger (wl_sk_), Oracul (vc_*_), npm, PyPI, RubyGems, GitLab, Telegram Bot, SendGrid, Shopify, DigitalOcean, Slack/Discord webhooks, Azure connections, GCP service accounts, SSH keys, JWTs, JDBC URLs, XML credentials, DB connections, emails, IPs, cards, and more.

Proxy + Chainwatch Chain (recommended for OpenClaw)

Stack both proxies for defense-in-depth:

OpenClaw → chainwatch:9999 (policy) → pastewatch:9998 (secrets) → api.anthropic.com

Setup as systemd services:

# 1. Pastewatch proxy (starts first)
cat > /etc/systemd/system/pastewatch-proxy.service << 'EOF'
[Unit]
Description=Pastewatch API Proxy (secret redaction)
After=network-online.target
Before=chainwatch-intercept.service
[Service]
Type=simple
ExecStart=/usr/local/bin/pastewatch-cli proxy \
  --port 9998 --upstream https://api.anthropic.com \
  --severity high --audit-log /var/log/pastewatch-proxy.log
Restart=always
RestartSec=3
MemoryMax=128M
[Install]
WantedBy=multi-user.target
EOF

# 2. Update chainwatch to forward to pastewatch (not Anthropic directly)
# Change --upstream from https://api.anthropic.com to http://localhost:9998

# 3. Enable and start
systemctl daemon-reload
systemctl enable pastewatch-proxy
systemctl start pastewatch-proxy
systemctl restart chainwatch-intercept

Rollback: Stop pastewatch-proxy, revert chainwatch --upstream to https://api.anthropic.com, restart chainwatch.

Guard — Block Secret-Leaking Commands

pastewatch-cli guard "cat .env"              # BLOCKED if .env has secrets
pastewatch-cli guard "psql -f migrate.sql"   # scans SQL file
pastewatch-cli guard "docker-compose up"     # scans env_files

Covers: shell builtins, DB CLIs (psql/mysql/mongosh/redis-cli/sqlite3), infra tools (ansible/terraform/docker/kubectl/helm), scripting (python/ruby/node/perl/php), file transfer (scp/rsync/ssh), pipe chains, subshells, redirects.

File Watcher

pastewatch-cli watch .              # continuous monitoring, real-time detection

Canary Tokens

pastewatch-cli canary generate --prefix myagent
pastewatch-cli canary verify
pastewatch-cli canary check --log /var/log/app.log

Encrypted Vault

pastewatch-cli --init-key                    # 256-bit key, mode 0600
pastewatch-cli fix --encrypt                 # → ChaCha20-Poly1305 vault
pastewatch-cli vault list | decrypt | export | rotate-key

Git History Scanning

pastewatch-cli scan --git-log
pastewatch-cli scan --git-log --range HEAD~50..HEAD --since 2025-01-01

Deduplicates by fingerprint. Output: text, json, sarif, markdown.

Org Posture Scanning

pastewatch-cli posture --org ppiankov           # scan all repos in org/user
pastewatch-cli posture --repos org/repo1,org/repo2
pastewatch-cli posture --compare previous.json  # trend tracking
pastewatch-cli posture --findings-only           # hide clean repos

Dashboard & Reports

pastewatch-cli dashboard                                  # aggregate across sessions
pastewatch-cli report --audit-log /var/log/pastewatch-audit.log
pastewatch-cli report --format json --since 2026-03-01T00:00:00Z

Configuration

pastewatch-cli init                         # generate .pastewatch.json
pastewatch-cli init --profile banking       # enterprise: JDBC, RFC 1918 IPs, service accounts
pastewatch-cli config                       # show resolved config
pastewatch-cli doctor                       # health check

Config cascade: /etc/pastewatch/config.json (admin) > project .pastewatch.json > ~/.config/pastewatch/config.json.

Features: sensitiveHosts, sensitiveIPPrefixes, xmlSensitiveTags, mcpMinSeverity, placeholderPrefix.

Detection Scope

30+ types: AWS keys/secrets, Anthropic/OpenAI/HuggingFace/Groq/Perplexity keys, DB connections, JDBC URLs, SSH keys, JWTs, emails, IPs, credit cards (Luhn), Slack/Discord webhooks, Azure, GCP service accounts, npm/PyPI/RubyGems/GitLab tokens, Telegram bot tokens, XML credentials, and more.

Gitignore-aware scanning (v0.21.0+). Deterministic regex. No ML. No API calls.

Limitations

  • Protects secrets from reaching LLM provider — does NOT protect prompt content or code structure
  • For full privacy, use a local model

Pastewatch MCP v1.2 Author: ppiankov Copyright © 2026 ppiankov Canonical source: https://github.com/ppiankov/pastewatch License: MIT

This tool follows the Agent-Native CLI Convention. Validate with: clawhub install ancc && ancc validate .

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

适合场景

01

OpenClaw 用户查找和安装 Skill 时

02

用户想查找某类 Agent Skill 时

03

需要根据任务场景推荐可安装能力包时

04

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

补充不同宿主或平台的使用分布数据

能力 5

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

OpenClaw

76.29%
按下载量换算4,212

安全审计

VirusTotal

可疑

ClawScan

可疑

Static analysis

通过

权限和风险

敏感数据

该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。来源安全扫描存在 warning/failed 结果,不能写成本站确认安全。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills