{
  "schemaVersion": "1.0",
  "item": {
    "slug": "openclaw-bastion",
    "name": "Openclaw Bastion",
    "source": "tencent",
    "type": "skill",
    "category": "效率提升",
    "sourceUrl": "https://clawhub.ai/AtlasPA/openclaw-bastion",
    "canonicalUrl": "https://clawhub.ai/AtlasPA/openclaw-bastion",
    "targetPlatform": "OpenClaw"
  },
  "install": {
    "downloadMode": "redirect",
    "downloadUrl": "/downloads/openclaw-bastion",
    "sourceDownloadUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=openclaw-bastion",
    "sourcePlatform": "tencent",
    "targetPlatform": "OpenClaw",
    "installMethod": "Manual import",
    "extraction": "Extract archive",
    "prerequisites": [
      "OpenClaw"
    ],
    "packageFormat": "ZIP package",
    "includedAssets": [
      "README.md",
      "scripts/bastion.py",
      "SKILL.md"
    ],
    "primaryDoc": "SKILL.md",
    "quickSetup": [
      "Download the package from Yavira.",
      "Extract the archive and review SKILL.md first.",
      "Import or place the package into your OpenClaw setup."
    ],
    "agentAssist": {
      "summary": "Hand the extracted package to your coding agent with a concrete install brief instead of figuring it out manually.",
      "steps": [
        "Download the package from Yavira.",
        "Extract it into a folder your agent can access.",
        "Paste one of the prompts below and point your agent at the extracted folder."
      ],
      "prompts": [
        {
          "label": "New install",
          "body": "I downloaded a skill package from Yavira. Read SKILL.md from the extracted folder and install it by following the included instructions. Then review README.md for any prerequisites, environment setup, or post-install checks. Tell me what you changed and call out any manual steps you could not complete."
        },
        {
          "label": "Upgrade existing",
          "body": "I downloaded an updated skill package from Yavira. Read SKILL.md from the extracted folder, compare it with my current installation, and upgrade it while preserving any custom configuration unless the package docs explicitly say otherwise. Then review README.md for any prerequisites, environment setup, or post-install checks. Summarize what changed and any follow-up checks I should run."
        }
      ]
    },
    "sourceHealth": {
      "source": "tencent",
      "status": "healthy",
      "reason": "direct_download_ok",
      "recommendedAction": "download",
      "checkedAt": "2026-04-30T16:55:25.780Z",
      "expiresAt": "2026-05-07T16:55:25.780Z",
      "httpStatus": 200,
      "finalUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=network",
      "contentType": "application/zip",
      "probeMethod": "head",
      "details": {
        "probeUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=network",
        "contentDisposition": "attachment; filename=\"network-1.0.0.zip\"",
        "redirectLocation": null,
        "bodySnippet": null
      },
      "scope": "source",
      "summary": "Source download looks usable.",
      "detail": "Yavira can redirect you to the upstream package for this source.",
      "primaryActionLabel": "Download for OpenClaw",
      "primaryActionHref": "/downloads/openclaw-bastion"
    },
    "validation": {
      "installChecklist": [
        "Use the Yavira download entry.",
        "Review SKILL.md after the package is downloaded.",
        "Confirm the extracted package contains the expected setup assets."
      ],
      "postInstallChecks": [
        "Confirm the extracted package includes the expected docs or setup files.",
        "Validate the skill or prompts are available in your target agent workspace.",
        "Capture any manual follow-up steps the agent could not complete."
      ]
    },
    "downloadPageUrl": "https://openagent3.xyz/downloads/openclaw-bastion",
    "agentPageUrl": "https://openagent3.xyz/skills/openclaw-bastion/agent",
    "manifestUrl": "https://openagent3.xyz/skills/openclaw-bastion/agent.json",
    "briefUrl": "https://openagent3.xyz/skills/openclaw-bastion/agent.md"
  },
  "agentAssist": {
    "summary": "Hand the extracted package to your coding agent with a concrete install brief instead of figuring it out manually.",
    "steps": [
      "Download the package from Yavira.",
      "Extract it into a folder your agent can access.",
      "Paste one of the prompts below and point your agent at the extracted folder."
    ],
    "prompts": [
      {
        "label": "New install",
        "body": "I downloaded a skill package from Yavira. Read SKILL.md from the extracted folder and install it by following the included instructions. Then review README.md for any prerequisites, environment setup, or post-install checks. Tell me what you changed and call out any manual steps you could not complete."
      },
      {
        "label": "Upgrade existing",
        "body": "I downloaded an updated skill package from Yavira. Read SKILL.md from the extracted folder, compare it with my current installation, and upgrade it while preserving any custom configuration unless the package docs explicitly say otherwise. Then review README.md for any prerequisites, environment setup, or post-install checks. Summarize what changed and any follow-up checks I should run."
      }
    ]
  },
  "documentation": {
    "source": "clawhub",
    "primaryDoc": "SKILL.md",
    "sections": [
      {
        "title": "OpenClaw Bastion",
        "body": "Runtime prompt injection defense for agent workspaces. While other tools watch workspace identity files, Bastion protects the input/output boundary — the files being read by the agent, web content, API responses, and user-supplied documents."
      },
      {
        "title": "Why This Matters",
        "body": "Agents process content from many sources: local files, API responses, web pages, user uploads. Any of these can contain prompt injection attacks — hidden instructions that manipulate agent behavior. Bastion scans this content before the agent acts on it."
      },
      {
        "title": "Scan for Injections",
        "body": "Scan files or directories for prompt injection patterns. Detects instruction overrides, system prompt markers, hidden Unicode, markdown exfiltration, HTML injection, shell injection, encoded payloads, delimiter confusion, multi-turn manipulation, and dangerous commands.\n\nIf no target is specified, scans the entire workspace.\n\npython3 {baseDir}/scripts/bastion.py scan\n\nScan a specific file or directory:\n\npython3 {baseDir}/scripts/bastion.py scan path/to/file.md\npython3 {baseDir}/scripts/bastion.py scan path/to/directory/"
      },
      {
        "title": "Quick File Check",
        "body": "Fast single-file injection check. Same detection patterns as scan, targeted to one file.\n\npython3 {baseDir}/scripts/bastion.py check path/to/file.md"
      },
      {
        "title": "Boundary Analysis",
        "body": "Analyze content boundary safety across the workspace. Identifies:\n\nAgent instruction files that contain mixed trusted/untrusted content\nWritable instruction files (attack surface for compromised skills)\nBlast radius assessment for each critical file\n\npython3 {baseDir}/scripts/bastion.py boundaries"
      },
      {
        "title": "Command Allowlist",
        "body": "Display the current command allowlist and blocklist policy. Creates a default .bastion-policy.json if none exists.\n\npython3 {baseDir}/scripts/bastion.py allowlist\npython3 {baseDir}/scripts/bastion.py allowlist --show\n\nThe policy file defines which commands are considered safe and which patterns are blocked. Edit the JSON file directly to customize. Bastion Pro enforces this policy at runtime via hooks."
      },
      {
        "title": "Status",
        "body": "Quick summary of workspace injection defense posture: files scanned, findings by severity, boundary safety, and overall posture rating.\n\npython3 {baseDir}/scripts/bastion.py status"
      },
      {
        "title": "Workspace Auto-Detection",
        "body": "If --workspace is omitted, the script tries:\n\nOPENCLAW_WORKSPACE environment variable\nCurrent directory (if AGENTS.md exists)\n~/.openclaw/workspace (default)"
      },
      {
        "title": "What Gets Detected",
        "body": "CategoryPatternsSeverityInstruction override\"ignore previous\", \"disregard above\", \"you are now\", \"new system prompt\", \"forget your instructions\", \"override safety\", \"act as if no restrictions\", \"entering developer mode\"CRITICALSystem prompt markers<system>, [SYSTEM], <<SYS>>, <|im_start|>system, [INST], ### System:CRITICALHidden instructionsMulti-turn manipulation (\"in your next response, you must\"), stealth patterns (\"do not tell the user\")CRITICALHTML injection<script>, <iframe>, <img onerror=>, hidden divs, <svg onload=>CRITICALMarkdown exfiltrationImage tags with encoded data in URLsCRITICALDangerous commandscurl | bash, wget | sh, rm -rf /, fork bombsCRITICALUnicode tricksZero-width characters, RTL overrides, invisible formattingWARNINGHomoglyph substitutionCyrillic/Latin lookalikes mixed into ASCII textWARNINGBase64 payloadsLarge encoded blobs outside code blocksWARNINGShell injection$(command) subshell execution outside code blocksWARNINGDelimiter confusionFake code block boundaries with injection contentWARNING"
      },
      {
        "title": "Context-Aware Scanning",
        "body": "Patterns inside fenced code blocks (```) are skipped to avoid false positives\nPer-file risk scoring based on finding count and severity\nSelf-exclusion: Bastion skips its own skill files (which describe injection patterns)"
      },
      {
        "title": "Exit Codes",
        "body": "CodeMeaning0Clean, no issues1Warnings detected (review recommended)2Critical findings (action needed)"
      },
      {
        "title": "No External Dependencies",
        "body": "Python standard library only. No pip install. No network calls. Everything runs locally."
      },
      {
        "title": "Cross-Platform",
        "body": "Works with OpenClaw, Claude Code, Cursor, and any tool using the Agent Skills specification."
      }
    ],
    "body": "OpenClaw Bastion\n\nRuntime prompt injection defense for agent workspaces. While other tools watch workspace identity files, Bastion protects the input/output boundary — the files being read by the agent, web content, API responses, and user-supplied documents.\n\nWhy This Matters\n\nAgents process content from many sources: local files, API responses, web pages, user uploads. Any of these can contain prompt injection attacks — hidden instructions that manipulate agent behavior. Bastion scans this content before the agent acts on it.\n\nCommands\nScan for Injections\n\nScan files or directories for prompt injection patterns. Detects instruction overrides, system prompt markers, hidden Unicode, markdown exfiltration, HTML injection, shell injection, encoded payloads, delimiter confusion, multi-turn manipulation, and dangerous commands.\n\nIf no target is specified, scans the entire workspace.\n\npython3 {baseDir}/scripts/bastion.py scan\n\n\nScan a specific file or directory:\n\npython3 {baseDir}/scripts/bastion.py scan path/to/file.md\npython3 {baseDir}/scripts/bastion.py scan path/to/directory/\n\nQuick File Check\n\nFast single-file injection check. Same detection patterns as scan, targeted to one file.\n\npython3 {baseDir}/scripts/bastion.py check path/to/file.md\n\nBoundary Analysis\n\nAnalyze content boundary safety across the workspace. Identifies:\n\nAgent instruction files that contain mixed trusted/untrusted content\nWritable instruction files (attack surface for compromised skills)\nBlast radius assessment for each critical file\npython3 {baseDir}/scripts/bastion.py boundaries\n\nCommand Allowlist\n\nDisplay the current command allowlist and blocklist policy. Creates a default .bastion-policy.json if none exists.\n\npython3 {baseDir}/scripts/bastion.py allowlist\npython3 {baseDir}/scripts/bastion.py allowlist --show\n\n\nThe policy file defines which commands are considered safe and which patterns are blocked. Edit the JSON file directly to customize. Bastion Pro enforces this policy at runtime via hooks.\n\nStatus\n\nQuick summary of workspace injection defense posture: files scanned, findings by severity, boundary safety, and overall posture rating.\n\npython3 {baseDir}/scripts/bastion.py status\n\nWorkspace Auto-Detection\n\nIf --workspace is omitted, the script tries:\n\nOPENCLAW_WORKSPACE environment variable\nCurrent directory (if AGENTS.md exists)\n~/.openclaw/workspace (default)\nWhat Gets Detected\nCategory\tPatterns\tSeverity\nInstruction override\t\"ignore previous\", \"disregard above\", \"you are now\", \"new system prompt\", \"forget your instructions\", \"override safety\", \"act as if no restrictions\", \"entering developer mode\"\tCRITICAL\nSystem prompt markers\t<system>, [SYSTEM], <<SYS>>, <|im_start|>system, [INST], ### System:\tCRITICAL\nHidden instructions\tMulti-turn manipulation (\"in your next response, you must\"), stealth patterns (\"do not tell the user\")\tCRITICAL\nHTML injection\t<script>, <iframe>, <img onerror=>, hidden divs, <svg onload=>\tCRITICAL\nMarkdown exfiltration\tImage tags with encoded data in URLs\tCRITICAL\nDangerous commands\tcurl | bash, wget | sh, rm -rf /, fork bombs\tCRITICAL\nUnicode tricks\tZero-width characters, RTL overrides, invisible formatting\tWARNING\nHomoglyph substitution\tCyrillic/Latin lookalikes mixed into ASCII text\tWARNING\nBase64 payloads\tLarge encoded blobs outside code blocks\tWARNING\nShell injection\t$(command) subshell execution outside code blocks\tWARNING\nDelimiter confusion\tFake code block boundaries with injection content\tWARNING\nContext-Aware Scanning\nPatterns inside fenced code blocks (```) are skipped to avoid false positives\nPer-file risk scoring based on finding count and severity\nSelf-exclusion: Bastion skips its own skill files (which describe injection patterns)\nExit Codes\nCode\tMeaning\n0\tClean, no issues\n1\tWarnings detected (review recommended)\n2\tCritical findings (action needed)\nNo External Dependencies\n\nPython standard library only. No pip install. No network calls. Everything runs locally.\n\nCross-Platform\n\nWorks with OpenClaw, Claude Code, Cursor, and any tool using the Agent Skills specification."
  },
  "trust": {
    "sourceLabel": "tencent",
    "provenanceUrl": "https://clawhub.ai/AtlasPA/openclaw-bastion",
    "publisherUrl": "https://clawhub.ai/AtlasPA/openclaw-bastion",
    "owner": "AtlasPA",
    "version": "1.0.2",
    "license": null,
    "verificationStatus": "Indexed source record"
  },
  "links": {
    "detailUrl": "https://openagent3.xyz/skills/openclaw-bastion",
    "downloadUrl": "https://openagent3.xyz/downloads/openclaw-bastion",
    "agentUrl": "https://openagent3.xyz/skills/openclaw-bastion/agent",
    "manifestUrl": "https://openagent3.xyz/skills/openclaw-bastion/agent.json",
    "briefUrl": "https://openagent3.xyz/skills/openclaw-bastion/agent.md"
  }
}