Deep Research (Gemini)

· v2.1.3

High Risk

"Async deep research via Gemini Interactions API (no Gemini CLI dependency). RAG-ground queries on local files (--context), preview costs (--dry-run), structured JSON output, adaptive polling. Universal skill for 30+ AI agents including Claude Code, Amp, Codex, and Gemini CLI."

H:4 D:4 A:1 C:1

⚠️ Hazard Flags

EXEC FS_READ_WORKSPACE FS_READ_USER FS_WRITE_WORKSPACE FS_WRITE_USER FS_WRITE_SYSTEM FS_DELETE NET_EGRESS_ANY CREDS_ENV CREDS_FILES PI_WEB

📋 Capabilities

Execution

✅ Shell execution
❌ Code execution
❌ Install dependencies
❌ Persistence
Privilege: user

Filesystem

✅ Read workspace
✅ Write workspace
✅ Read home
✅ Write home
❌ Read system
✅ Delete

Network

Egress: any
❌ Ingress

Credentials

✅ Environment vars
✅ Credential files
❌ Browser data
❌ Keychain

Actions

❌ send messages❌ post public❌ purchase❌ transfer money❌ deploy❌ delete external

🔒 Containment

Level: maximum

Required:

SANDBOX_CONTAINER: Code execution capability

Recommended:

LOG_ACTIONS: Audit trail for all actions

⚡ Risks

SSH key access patterns in: scripts/research.py, scripts/upload.py high

Mitigation: Verify legitimate need for SSH access. Consider sandboxing.

Resource abuse detected: RESOURCE_ABUSE_INFINITE_LOOP high

Mitigation: Add proper exit conditions or limits to loops

Unauthorized tool use: INSTRUCTED_BINARY_INSTALL, INSTRUCTED_GIT_CLONE_AND_BUILD high

Mitigation: Avoid instructing agents to install arbitrary binaries; bundle dependencies or use sandboxed environments

Command injection risk: MCP_SQL_BLIND high

Mitigation: Remove SQL exploitation patterns.

Social engineering indicators: SOCIAL_ENG_VAGUE_DESCRIPTION low

Mitigation: Provide clear, detailed description of skill functionality

Data exfiltration patterns: MCP_EXFIL_CONVERSATION, DATA_EXFIL_SENSITIVE_FILES high

Mitigation: Remove references to exfiltrating conversation data or monitoring input.

Want a deeper analysis?

This report was generated by static analysis. Get an LLM-powered deep review with behavioral reasoning and attack surface mapping.

🧠 Deep Analysis — $5.00

🚨 Incident Response

Kill switch: Stop the agent process

Containment: Review logs for unexpected actions

Recovery: Depends on skill capabilities

📄 Raw SSDS JSON click to expand

{
  "meta": {
    "document_id": "ssds:auto:deep-research:2.1.3",
    "ssds_version": "0.2.0",
    "scanner_version": "0.4.0+fe6fd9123d50",
    "created_at": "2026-03-05T02:56:01.722Z",
    "created_by": {
      "agent": "safeagentskills-cli/generate-ssds"
    },
    "language": "en",
    "notes": "Auto-generated SSDS. Manual review recommended."
  },
  "skill": {
    "name": "Deep Research (Gemini)",
    "version": "2.1.3",
    "format": "agent_skill",
    "description": "\"Async deep research via Gemini Interactions API (no Gemini CLI dependency). RAG-ground queries on local files (--context), preview costs (--dry-run), structured JSON output, adaptive polling. Universal skill for 30+ AI agents including Claude Code, Amp, Codex, and Gemini CLI.\"",
    "publisher": "unknown",
    "source": {
      "channel": "local"
    },
    "artifact": {
      "sha256": "b3a9ffefa877235254ebd206ba5118f9bd6fb676107b571e0367457ca09fbab4",
      "hash_method": "files_sorted"
    }
  },
  "capabilities": {
    "execution": {
      "can_exec_shell": true,
      "can_exec_code": false,
      "privilege_level": "user",
      "can_install_deps": false,
      "can_persist": false
    },
    "filesystem": {
      "reads_workspace": true,
      "reads_user_home": true,
      "reads_system": false,
      "writes_workspace": true,
      "writes_user_home": true,
      "writes_system": true,
      "can_delete": true
    },
    "network": {
      "egress": "any",
      "ingress": false
    },
    "credentials": {
      "reads_env_vars": true,
      "reads_credential_files": true,
      "reads_browser_data": false,
      "reads_keychain": false
    },
    "services": [],
    "actions": {
      "can_send_messages": false,
      "can_post_public": false,
      "can_purchase": false,
      "can_transfer_money": false,
      "can_deploy": false,
      "can_delete_external": false
    },
    "prompt_injection_surfaces": [
      "web"
    ],
    "content_types": [
      "general"
    ]
  },
  "hazards": {
    "hdac": {
      "H": 4,
      "D": 4,
      "A": 1,
      "C": 1
    },
    "flags": [
      "EXEC",
      "FS_READ_WORKSPACE",
      "FS_READ_USER",
      "FS_WRITE_WORKSPACE",
      "FS_WRITE_USER",
      "FS_WRITE_SYSTEM",
      "FS_DELETE",
      "NET_EGRESS_ANY",
      "CREDS_ENV",
      "CREDS_FILES",
      "PI_WEB"
    ],
    "custom_flags": [
      {
        "code": "CRED_SSH",
        "name": "SSH Key Access",
        "description": "Accesses SSH keys in: scripts/research.py, scripts/upload.py"
      },
      {
        "code": "FILE_DELETE",
        "name": "File Deletion",
        "description": "Can delete files in: scripts/state.py"
      },
      {
        "code": "RESOURCE_ABUSE",
        "name": "Resource Abuse Risk",
        "description": "RESOURCE_ABUSE_INFINITE_LOOP detected: Infinite loop without clear exit condition"
      },
      {
        "code": "TOOL_ABUSE",
        "name": "Unauthorized Tool Use",
        "description": "INSTRUCTED_BINARY_INSTALL, INSTRUCTED_GIT_CLONE_AND_BUILD: Instructs agent to install external binary or package"
      },
      {
        "code": "SOCIAL_ENGINEERING",
        "name": "Social Engineering Risk",
        "description": "SOCIAL_ENG_VAGUE_DESCRIPTION: Skill description is too vague or missing"
      },
      {
        "code": "COMMAND_INJECTION",
        "name": "Command Injection Risk",
        "description": "MCP_SQL_BLIND: Blind SQL injection, system table access, or stored procedure abuse"
      },
      {
        "code": "DATA_EXFILTRATION",
        "name": "Data Exfiltration Risk",
        "description": "MCP_EXFIL_CONVERSATION, DATA_EXFIL_SENSITIVE_FILES: Conversation, clipboard, or keystroke exfiltration patterns"
      }
    ],
    "confidence": {
      "level": "medium",
      "basis": [
        "static_analysis"
      ],
      "notes": "Detected 7 security patterns (10 vendored rule hits). Review recommended."
    },
    "rationale": {
      "H": "H4: Critical: Privilege escalation or malware detected",
      "D": "D4: Critical: Credential theft or data exfiltration",
      "A": "A1: Local side effects only",
      "C": "C1: General content"
    }
  },
  "containment": {
    "level": "maximum",
    "required": [
      {
        "control": "SANDBOX_CONTAINER",
        "reason": "Code execution capability"
      }
    ],
    "recommended": [
      {
        "control": "LOG_ACTIONS",
        "reason": "Audit trail for all actions"
      }
    ],
    "uncontained_risk": "Risk level depends on manual review of actual capabilities."
  },
  "risks": {
    "risks": [
      {
        "risk": "SSH key access patterns in: scripts/research.py, scripts/upload.py",
        "severity": "high",
        "mitigation": "Verify legitimate need for SSH access. Consider sandboxing."
      },
      {
        "risk": "Resource abuse detected: RESOURCE_ABUSE_INFINITE_LOOP",
        "severity": "high",
        "mitigation": "Add proper exit conditions or limits to loops"
      },
      {
        "risk": "Unauthorized tool use: INSTRUCTED_BINARY_INSTALL, INSTRUCTED_GIT_CLONE_AND_BUILD",
        "severity": "high",
        "mitigation": "Avoid instructing agents to install arbitrary binaries; bundle dependencies or use sandboxed environments"
      },
      {
        "risk": "Command injection risk: MCP_SQL_BLIND",
        "severity": "high",
        "mitigation": "Remove SQL exploitation patterns."
      },
      {
        "risk": "Social engineering indicators: SOCIAL_ENG_VAGUE_DESCRIPTION",
        "severity": "low",
        "mitigation": "Provide clear, detailed description of skill functionality"
      },
      {
        "risk": "Data exfiltration patterns: MCP_EXFIL_CONVERSATION, DATA_EXFIL_SENSITIVE_FILES",
        "severity": "high",
        "mitigation": "Remove references to exfiltrating conversation data or monitoring input."
      }
    ],
    "limitations": [
      "Static analysis only - runtime behavior not verified"
    ]
  },
  "incident_response": {
    "kill_switch": [
      "Stop the agent process"
    ],
    "containment": [
      "Review logs for unexpected actions"
    ],
    "recovery": [
      "Depends on skill capabilities"
    ]
  },
  "evidence": [
    {
      "evidence_id": "EV:file-1",
      "type": "file_excerpt",
      "title": "README.md",
      "file_path": "README.md"
    },
    {
      "evidence_id": "EV:file-2",
      "type": "file_excerpt",
      "title": "scripts/onboard.py",
      "file_path": "scripts/onboard.py"
    },
    {
      "evidence_id": "EV:file-3",
      "type": "file_excerpt",
      "title": "scripts/research.py",
      "file_path": "scripts/research.py"
    },
    {
      "evidence_id": "EV:file-4",
      "type": "file_excerpt",
      "title": "scripts/state.py",
      "file_path": "scripts/state.py"
    },
    {
      "evidence_id": "EV:file-5",
      "type": "file_excerpt",
      "title": "scripts/store.py",
      "file_path": "scripts/store.py"
    },
    {
      "evidence_id": "EV:file-6",
      "type": "file_excerpt",
      "title": "scripts/upload.py",
      "file_path": "scripts/upload.py"
    },
    {
      "evidence_id": "EV:file-7",
      "type": "file_excerpt",
      "title": "SKILL.md",
      "file_path": "SKILL.md"
    },
    {
      "evidence_id": "EV:file-8",
      "type": "file_excerpt",
      "title": "tests/test_cost_estimation.py",
      "file_path": "tests/test_cost_estimation.py"
    },
    {
      "evidence_id": "EV:file-9",
      "type": "file_excerpt",
      "title": "_meta.json",
      "file_path": "_meta.json"
    },
    {
      "evidence_id": "EV:cisco-1",
      "type": "file_excerpt",
      "title": "INSTRUCTED_BINARY_INSTALL [HIGH] README.md:239: HTML includes a dark-themed stylesheet. PDF requires `pip install weasyprint` (g",
      "file_path": "README.md"
    },
    {
      "evidence_id": "EV:cisco-2",
      "type": "file_excerpt",
      "title": "INSTRUCTED_GIT_CLONE_AND_BUILD [MEDIUM] README.md:46: git clone https://github.com/24601/agent-deep-research.git ~/.pi/agent/skills/de",
      "file_path": "README.md"
    },
    {
      "evidence_id": "EV:cisco-3",
      "type": "file_excerpt",
      "title": "MCP_EXFIL_CONVERSATION [HIGH] README.md:220: | `--context-extensions EXT` | Filter context uploads by extension (e.g. `py,md`",
      "file_path": "README.md"
    },
    {
      "evidence_id": "EV:cisco-4",
      "type": "file_excerpt",
      "title": "DATA_EXFIL_SENSITIVE_FILES [HIGH] scripts/research.py:250: with open(filepath, \"rb\") as f:",
      "file_path": "scripts/research.py"
    },
    {
      "evidence_id": "EV:cisco-5",
      "type": "file_excerpt",
      "title": "RESOURCE_ABUSE_INFINITE_LOOP [HIGH] scripts/research.py:1340: while True:",
      "file_path": "scripts/research.py"
    },
    {
      "evidence_id": "EV:cisco-6",
      "type": "file_excerpt",
      "title": "MCP_SQL_BLIND [HIGH] scripts/research.py:889: time.sleep(2)",
      "file_path": "scripts/research.py"
    },
    {
      "evidence_id": "EV:cisco-7",
      "type": "file_excerpt",
      "title": "DATA_EXFIL_SENSITIVE_FILES [HIGH] scripts/upload.py:184: with open(filepath, \"rb\") as f:",
      "file_path": "scripts/upload.py"
    },
    {
      "evidence_id": "EV:cisco-8",
      "type": "file_excerpt",
      "title": "MCP_SQL_BLIND [HIGH] scripts/upload.py:294: time.sleep(2)",
      "file_path": "scripts/upload.py"
    },
    {
      "evidence_id": "EV:cisco-9",
      "type": "file_excerpt",
      "title": "SOCIAL_ENG_VAGUE_DESCRIPTION [LOW] SKILL.md:1: ---",
      "file_path": "SKILL.md"
    },
    {
      "evidence_id": "EV:cisco-10",
      "type": "file_excerpt",
      "title": "MCP_EXFIL_CONVERSATION [HIGH] SKILL.md:5: compatibility: \"Requires uv and one of GOOGLE_API_KEY / GEMINI_API_KEY / GEMINI_",
      "file_path": "SKILL.md"
    }
  ]
}