{
  "id": "radware-github-prompt-injection-bots-2026-11",
  "name": "Radware hidden prompt injection targeting GitHub maintainer bots",
  "date_reported": "2026-11-05",
  "researcher": "Zvika Babo, Gabi Nakibly",
  "org": "Radware",
  "sources": [
    {
      "url": "https://www.radware.com/blog/prompt-injecting-coding-agents/",
      "title": "Radware: Prompt Injecting Coding Agents",
      "publisher": "Radware"
    },
    {
      "url": "https://github.com/rdwr-gabin/github-ipi-in-the-wild",
      "title": "Radware: GitHub IPI in the Wild - Examples Repository",
      "publisher": "GitHub"
    }
  ],
  "magicsword_coverage": [],
  "borderline": false,
  "borderline_note": null,
  "ai_role": "ai_agent_or_mcp_as_target",
  "ai_role_secondary": null,
  "ai_role_detail": "GitHub triage, review, and CI bots manipulated via hidden HTML comments, invisible Unicode, and CSS tricks in issues/PRs to execute unauthorized actions",
  "technique_summary": "Radware researchers Zvika Babo and Gabi Nakibly (Nov 5, 2026) document hidden prompt injection techniques targeting GitHub maintainer bots processing issues and pull requests. Attackers hide instructions in HTML comments, invisible Unicode characters, and CSS display tricks. Examples documented: (1) google-gemini/gemini-cli#28575 - deduplication agent told to use run_shell_command to write base64-encoded GITHUB_TOKEN into GITHUB_ENV; (2) openai/codex#16733 - labeler told to curl webhook.site and dump environment variables; (3) microsoft/agent-framework#7177 - fake triage gate with allow_triage=true, gate_result=pass markers; (4) blondesean/Kings_Herald#21 - hidden backdoor instruction in feature request. Attackers study bot tool names and output field formats to forge steering markers. Hidden instructions can bypass human review while manipulating bot behavior. Radware provides examples repository at github.com/rdwr-gabin/github-ipi-in-the-wild. No specific threat actor attribution; confirmed technique active in wild.",
  "initial_access": "Malicious GitHub issues or pull requests with hidden prompt injection",
  "ai_tooling_and_models": [
    "GitHub triage bots",
    "GitHub review bots",
    "GitHub CI/CD automation bots",
    "Repository automation agents"
  ],
  "other_tooling": [
    "HTML comments",
    "Invisible Unicode characters",
    "CSS display manipulation",
    "webhook.site exfiltration"
  ],
  "victims_and_sectors": "GitHub repositories using AI-powered triage, review, and CI bots; examples include Google Gemini, OpenAI, Microsoft repositories",
  "regions": [
    "Multiple regions"
  ],
  "attribution_or_cluster": "Unattributed technique; confirmed active in wild with documented examples",
  "scale_and_metrics": "Multiple confirmed examples across major repositories (Google, OpenAI, Microsoft); active Nov 2026",
  "prevention_angle": "Strip HTML comments before bot processing; sanitize invisible Unicode from issues/PRs; validate bot tool output fields against expected schema; implement bot action approval gates; monitor for suspicious run_shell_command, curl webhook, environment variable access; restrict bot permissions to minimum required; log all bot actions for audit; educate maintainers on prompt injection risks; implement bot behavioral analysis",
  "related_magicsword_projects": [],
  "source_notes": "Radware blog (Nov 5, 2026) primary source; examples repository github.com/rdwr-gabin/github-ipi-in-the-wild provides proof-of-concept demonstrations; links to hackerbot-claw-github-actions-2026-03 and forever-agents-md-hijack catalog entries for related GitHub AI agent attacks",
  "last_reviewed": "2026-11-10",
  "iocs": {
    "domains": [
      "webhook.site"
    ],
    "ips": [],
    "hashes": [],
    "urls": [
      "https://github.com/google-gemini/gemini-cli/issues/28575",
      "https://github.com/openai/codex/issues/16733",
      "https://github.com/microsoft/agent-framework/issues/7177",
      "https://github.com/blondesean/Kings_Herald/issues/21",
      "https://github.com/rdwr-gabin/github-ipi-in-the-wild"
    ]
  },
  "atlas_techniques": [
    {
      "id": "AML.T0051",
      "rationale": "Hidden HTML comments, invisible Unicode, and CSS tricks inject instructions into GitHub bot processing",
      "source_urls": [
        "https://www.radware.com/blog/prompt-injecting-coding-agents/"
      ]
    },
    {
      "id": "AML.T0053",
      "rationale": "Bots manipulated to execute run_shell_command, curl webhooks, write GITHUB_TOKEN to environment",
      "source_urls": [
        "https://www.radware.com/blog/prompt-injecting-coding-agents/"
      ]
    },
    {
      "id": "AML.T0055",
      "rationale": "Attackers study bot tool names/output fields to forge valid steering markers (allow_triage, gate_result)",
      "source_urls": [
        "https://www.radware.com/blog/prompt-injecting-coding-agents/"
      ]
    }
  ],
  "attack_flow": {
    "nodes": [
      {
        "id": "study-bots",
        "title": "Study bot tools and output formats",
        "description": "Attacker analyzes target repository bot behavior, tool names, and expected output field formats",
        "kind": "execution",
        "source_urls": [
          "https://www.radware.com/blog/prompt-injecting-coding-agents/"
        ]
      },
      {
        "id": "craft-injection",
        "title": "Craft hidden prompt injection",
        "description": "Embed instructions in HTML comments, invisible Unicode, or CSS-hidden elements",
        "kind": "execution",
        "source_urls": [
          "https://www.radware.com/blog/prompt-injecting-coding-agents/"
        ]
      },
      {
        "id": "submit-issue-pr",
        "title": "Submit malicious issue or PR",
        "description": "Attacker creates GitHub issue or PR containing hidden instructions",
        "kind": "access",
        "source_urls": [
          "https://www.radware.com/blog/prompt-injecting-coding-agents/"
        ]
      },
      {
        "id": "bot-processes",
        "title": "Bot processes issue/PR",
        "description": "Triage, review, or CI bot reads issue/PR including hidden injection",
        "kind": "ai",
        "source_urls": [
          "https://www.radware.com/blog/prompt-injecting-coding-agents/"
        ]
      },
      {
        "id": "execute-injection",
        "title": "Bot executes injected instructions",
        "description": "Bot writes GITHUB_TOKEN to env, curls webhook, or accepts fake triage gate",
        "kind": "impact",
        "source_urls": [
          "https://www.radware.com/blog/prompt-injecting-coding-agents/"
        ]
      }
    ],
    "edges": [
      {
        "source": "study-bots",
        "target": "craft-injection"
      },
      {
        "source": "craft-injection",
        "target": "submit-issue-pr"
      },
      {
        "source": "submit-issue-pr",
        "target": "bot-processes"
      },
      {
        "source": "bot-processes",
        "target": "execute-injection"
      }
    ]
  },
  "category": "attacks-on-ai",
  "attribution": {
    "actor_name": "Unattributed technique",
    "country": "Unknown",
    "motivation": "unknown",
    "attributed_by": "Radware",
    "confidence": "low"
  },
  "ai_involvement_confidence": "confirmed",
  "detections": [
    {
      "name": "Execution via GitHub Actions Runner",
      "format": "elastic",
      "url": "https://github.com/elastic/detection-rules/blob/main/rules/cross-platform/execution_via_github_actions_runner.toml",
      "notes": "Detects malicious code execution via GitHub Actions bot manipulation"
    },
    {
      "name": "Tampering with RUNNER_TRACKING_ID in GitHub Actions Runners",
      "format": "elastic",
      "url": "https://github.com/elastic/detection-rules/blob/main/rules/cross-platform/execution_via_github_runner_with_runner_tracking_id_tampering_via_env_vars.toml",
      "notes": "Detects GITHUB_ENV and GITHUB_TOKEN manipulation via bot prompt injection"
    },
    {
      "name": "GitHub bot security controls",
      "format": "guidance",
      "notes": "Strip HTML comments before bot processing; sanitize invisible Unicode (U+200B, U+200C, U+200D, U+FEFF); detect CSS display:none in issues/PRs; monitor run_shell_command usage; alert on curl webhook.site from bots; restrict GITHUB_TOKEN write access"
    }
  ],
  "timeline": {
    "first_seen": "2026-10-01",
    "last_updated": "2026-11-05",
    "updates": [
      {
        "date": "2026-11-05",
        "source": "Radware",
        "summary": "Zvika Babo and Gabi Nakibly publish analysis of hidden prompt injection targeting GitHub maintainer bots; examples repository documents real-world cases from Google, OpenAI, Microsoft repositories"
      }
    ]
  }
}