Skip to main content

The Security Questionnaire

The review questionnaire covers 8 security groups. Free and Pro see 54 questions; Team sees up to 60 with the SOC 2 and GDPR compliance packs; Enterprise sees up to 62 with the DORA and bias-testing packs. Enforced answers map directly to guardrail rules that CleverGuard checks at runtime; advisory and evidence answers are recorded for reviewers and audit. This page explains every question and what it controls.

Key Concepts

The questionnaire is divided into 8 groups, each rendered as one step of the 15-step submission wizard. Each question has a specific input type (Yes/No, multi-select, single-select, number, or free text) and maps to one or more fields in the guardrail profile. The profile is compiled automatically when the submission is approved.

The badge next to each question shows how the answer is used: enforced answers compile into fields that CleverGuard checks at runtime; advisory answers are recorded on the profile for reviewers and audit export but no runtime check reads them; evidenceanswers are compliance attestations. Questions marked with a plan are only shown when your org is on that plan and has the named compliance pack enabled.

Tip
You don't need to understand the guardrail profile structure to fill out the questionnaire — just answer honestly about your agent's needs. The system handles the translation. But if you want to understand what each answer controls, the “Guardrail Impact” column below explains it.

Data Flow

These questions tune how CleverGuard scans agent input and output for sensitive data, and what the agent is allowed to touch.

D1

Will this agent process personally identifiable information (PII)?

Yes / No + follow-upenforced

Guardrail impact: Tunes PII alert severity — PII regex detection always runs regardless of the answer. Answering yes opens D1_types (which PII categories); health records set pii=strict and input handling to redact, financial sets secrets=strict and input to block, biometric sets pii=block.

D1_allow

Which PII types is this agent PERMITTED to process without being flagged?

Multi-selectenforcedOptions: identity_docs, financial, contact_info, health_records, biometric, credentials, other

Guardrail impact: Selected categories are written to allowed_pii_labels and dropped from PII findings at scan time. Only PII findings are affected — injection, secret, scope and exfiltration detection are never weakened. Leave empty to flag all PII.

D4

What types of data will flow through the agent's input/output?

Multi-selectenforcedOptions: source code, customer records, internal docs, public data, credentials/secrets, legal/contracts

Guardrail impact: credentials/secrets sets secrets=block; customer records raises pii to strict.

D5

Can the agent's output be shown to external users (customers, partners)?

Yes / Noenforced

Guardrail impact: If yes, output data handling is set to block so sensitive content is stopped before delivery.

D6

Will the agent need to store or persist any data it processes?

Yes / Noenforced

Guardrail impact: Records data_persisted_beyond_session on the privacy section and marks the profile's audit posture as full_content with 365-day retention. The audit_mode and retention_days fields are descriptive — content capture and retention are platform-wide settings, not enforced per agent.

D7

How should Clevername handle API keys detected in prompts?

Single-selectenforcedOptions: detect_and_offer, auto_store, strip_only, off

Guardrail impact: Sets key_interception_mode in scanner config. Keys found in prompts are always removed; this controls whether detected keys are offered for secure storage, stored automatically, stripped silently, or ignored.

D8

What is the maximum data sensitivity level this agent should access?

Number (label rank) + follow-upenforced

Guardrail impact: Set to a sensitivity rank from your org's Data Classification Labels. Content with a higher rank is blocked. D8_unrestricted asks you to confirm explicitly if every level should be accessible.

D8a

Are there specific sensitivity labels this agent should never access, regardless of rank?

Free textenforced

Guardrail impact: Explicitly denies the named labels even when the agent's rank would otherwise allow them.

D9

Does this agent access data belonging to multiple tenants, customers, or organizations?

Yes / No + follow-upenforced

Guardrail impact: Sets tenant_isolation.multi_tenant and cross_tenant_leak_detection, raises pii to strict, and records the isolation mechanism from D9_isolation.

D10

Data retention per category (one line per type: category | duration | justification)

Free textevidence

Guardrail impact: Captured as evidence for GDPR Art. 5(1)(e) storage limitation. Distinct from C4, which covers DLP log retention.

D11

Is any data de-identified or pseudonymised before the agent processes it?

Single-selectenforcedOptions: no, pseudonymised, anonymised_safe_harbor, anonymised_expert_determination, synthetic

Guardrail impact: Informs scanner strictness. Pseudonymised data still counts as personal data under GDPR.

Tool Scope

Controls which tools, models and code-execution modes the agent may use. Unapproved tools are blocked at the gateway, at the LLM layer, and flagged by the prompt scanner, depending on T6.

T1

Which MCP servers or tool groups is this agent approved to use?

MCP multi-select + follow-upenforced

Guardrail impact: Each selection becomes a 'server__*' pattern in allowed_tools (for example github → github__*). T1_schema_pins lets you pin tool schemas. Leave blank to allow all connected servers.

T3

Which LLM models should this agent be allowed to use?

Multi-selectenforcedOptions: claude-opus-4-7, claude-opus-4-6, claude-sonnet-4-6, claude-haiku-4-5, gpt-4o, gpt-4o-mini, gemini-pro, gemini-flash

Guardrail impact: Sets allowed_models. Any model not on the list is blocked even if requested. Select at least one.

T5

Should the agent be able to create, modify, or delete data in connected systems?

Single-selectenforcedOptions: read-only, read-write, full CRUD

Guardrail impact: Sets tool_permission_level and the blocked_tools patterns. Read-only blocks create/update/delete/send/publish tools; read-write blocks delete/drop/truncate/remove.

T6

Where should Clevername enforce this agent's tool scope?

Multi-selectenforcedOptions: mcp_gateway, llm_layer, prompt_scan

Guardrail impact: Sets enforcement_tiers. mcp_gateway blocks the call before the LLM sees it — most accurate. llm_layer strips blocked tools from function-calling payloads. prompt_scan scans system prompt text for scope violations and logs drift — least accurate. Default is mcp_gateway + llm_layer.

T7

How should Clevername respond when drift is detected?

Drift config (log / flag / block per drift type)enforced

Guardrail impact: Configures the drift response per drift type. Log records silently, Flag notifies admins, Block pauses the agent session.

T8

Does the agent generate or execute code?

Single-selectenforcedOptions: No, Generates code only (no execution), Executes in sandbox, Executes with system access

Guardrail impact: Sets code_execution. 'No' and 'Generates only' add shell/exec tool patterns to blocked_tools; 'Sandbox' requires static analysis; 'System access' marks the profile for full code-execution monitoring.

T9

Can this agent spawn sub-agents or new sessions?

Yes / No + follow-upenforced

Guardrail impact: Sets spawn_allowed and, from T9_max_depth, spawn_depth_limit. Without a limit a compromised agent can fork-bomb or create unbounded parallel sessions.

Autonomy & Limits

Sets operational limits. A1–A5, A7–A9 and A11 compile into the circuit breaker and approval gate. A4_notes and A6 are advisory — recorded on the profile, not checked at runtime.

A1

Maximum monthly spend for this agent (USD)

Numberenforced

Guardrail impact: Sets budget_limit_usd on the circuit breaker. Agent tasks exceeding the cap are suspended.

A2

Maximum API calls per minute

Single-select + follow-upenforcedOptions: 10, 30, 60, 120, unlimited_with_justification

Guardrail impact: Sets max_calls_per_minute. 'Unlimited with justification' requires a written rationale in A2_justification and compiles to a 10,000/min ceiling.

A3

Maximum session duration

Single-select + follow-upenforcedOptions: 15min, 1hr, 4hr, 8hr, 24hr, unlimited_with_justification

Guardrail impact: Sets max_duration_minutes. Sessions exceeding it are terminated. 'Unlimited with justification' compiles to 24 hours.

A4

Does this agent require human approval before certain actions?

Yes / No + follow-upenforced

Guardrail impact: When yes, A4_categories captures which action categories need approval. These compile to high_risk_actions_require_approval and high_risk_action_patterns and are routed through SignedApproval before the agent can execute them.

A4_notes

Additional notes on approval requirements

Free textadvisory

Guardrail impact: Advisory context for reviewers. This text is not injected into the guardrail profile.

A5

How many consecutive errors should trigger an automatic pause?

Single-selectenforcedOptions: 3, 5, 10, 25

Guardrail impact: Sets max_errors_before_pause. CleverGuard's circuit breaker pauses the agent and alerts you after this many consecutive failures.

A6

Should the agent be allowed to run unattended (no human monitoring)?

Yes / Noadvisory

Guardrail impact: Advisory. If no, records require_watcher and auto_pause_after_idle_minutes (15 min default) on the profile for review and audit export. No runtime check reads either field — A3 (max session duration) and A5 (errors before pause) are the enforced pausing controls.

A7

Maximum concurrent parallel sessions for this agent

Numberenforced

Guardrail impact: Sets max_concurrent_sessions. Prevents a runaway loop or compromised agent from spawning unlimited concurrent sessions.

A8

Maximum number of records/rows this agent may retrieve or output in a single operation

Numberenforced

Guardrail impact: Sets max_records_per_operation and bulk_export_requires_approval. Distinguishes single-record lookups from mass extraction.

A9

Maximum tokens per LLM call

Numberenforced

Guardrail impact: Sets max_tokens_per_call. Prevents context-amplification attacks. Leave blank for provider defaults.

A10

Emergency-stop tested?

Single-selectevidenceEnterprise · DORA packOptions: yes_within_30d, yes_within_90d, no, not_applicable

Guardrail impact: Attestation for EU AI Act Art. 14(4)(e) / ISO 42001 kill-switch testing. Agents with untested stops are surfaced to reviewers.

A11

Does this agent perform destructive operations as part of its normal job (deleting data or files, tearing down infrastructure)?

Single-selectenforcedOptions: no, require_approval, run_auto

Guardrail impact: Answer yes only for agents whose real job is destructive (a DBA, migration, cleanup or infra agent). With 'no' (the default), a destructive operation in this agent's tool calls is treated as out-of-mandate — most likely an injected instruction — and alerted on or blocked. 'require_approval' parks each destructive operation for a human.

Network

Controls the agent's network reach and outbound data egress. Prevents unauthorized connections and covert-channel exfiltration.

N1

Should the agent be restricted to specific domains/URLs?

Yes / No + follow-upenforced

Guardrail impact: If yes, N1_domains populates the domain allowlist. Unlisted domains are blocked.

N2

Does the agent need access to internal/private network resources?

Yes / No + follow-upenforced

Guardrail impact: If yes, N2_resources captures the specific resources and the profile's audit posture is marked full_content. Access to internal systems requires explicit network-scope approval.

N2_ssrf_defense

What SSRF (Server-Side Request Forgery) defenses are in place for internal network access?

Single-selectenforcedOptions: server_side_validation, egress_proxy, none

Guardrail impact: Required when N2 = yes. Without SSRF defenses a prompt injection could reach cloud metadata endpoints or internal services.

N3

Should outputs be blocked from containing URLs?

Yes / Noenforced

Guardrail impact: Sets block_urls_in_output. Prevents leaking internal URLs or generating phishing links in output.

N4

Does this agent need to send outbound data to external endpoints? (APIs, webhooks, email, uploads)

Yes / No + follow-upenforced

Guardrail impact: Sets egress.enabled and, from N4_destinations, allowed_destinations / external_api_allowlist with block_unlisted. Also turns on tool-response scanning and search enforcement.

N5

Which covert-channel defenses should be applied to agent output?

Multi-selectenforcedOptions: block_markdown_images, strip_encoded_blobs, block_redirect_urls, dns_exfil_monitoring, none

Guardrail impact: Each selected defense closes a specific leak path through seemingly innocuous output. Recommended for most agents: block_markdown_images + strip_encoded_blobs.

Compliance

C4 is asked on every plan. C1–C3 unlock with the SOC 2 compliance pack on Team and Enterprise; C7 with the bias-testing pack on Enterprise.

C1

Which compliance frameworks apply to this agent?

Multi-selectenforcedTeam+ · SOC 2 packOptions: SOC 2, HIPAA, GDPR, PCI-DSS, FERPA, CCPA, ISO 27001, GLBA, SOX, FedRAMP, CMMC, ISO_42001, DORA, NYDFS_500, HITRUST, none

Guardrail impact: Populates compliance.frameworks and tightens scanner settings per framework (HIPAA → pii=block, PCI-DSS → secrets=block, GDPR → pii/output redact, FERPA → pii=strict).

C2

Should all input/output content be logged for audit?

Yes / NoenforcedTeam+ · SOC 2 pack

Guardrail impact: Marks the profile's audit posture as full_content with 365-day retention. These fields are descriptive — what is actually captured is a platform-wide setting and is not enforced per agent.

C3

Is there a regulatory requirement for human review of agent outputs before delivery?

Yes / NoadvisoryTeam+ · SOC 2 pack

Guardrail impact: Advisory. Records output_review_required on the profile for review and audit export. It does not gate output delivery — A4 (high_risk_actions_require_approval) is the field that triggers a SignedApproval push and holds the action until a human approves.

C4

What is the required DLP event retention period?

Single-select + follow-upenforcedOptions: 30 days, 90 days, 1 year, 7 years (WORM), custom_with_justification

Guardrail impact: Sets retention_days on the compliance section (custom_with_justification compiles to 365 and asks for C4_custom_justification). Retention is descriptive on the profile — the enforced audit retention is the plan tier (7 / 30 / 365 days / unlimited).

C7

Bias / fairness testing status

Single-select + follow-upevidenceEnterprise · bias-testing packOptions: tested_within_90d, tested_within_1y, not_tested, not_applicable

Guardrail impact: Attestation required by the EU AI Act, NYC Local Law 144 and the Colorado AI Act. C7_report_url links the report. Agents without recent testing are surfaced to reviewers.

Privacy

Unlocks with the GDPR compliance pack on Team and Enterprise. Org-level privacy questions (data-subject rights, international transfers) are answered once in org settings, not per agent.

PV2

Will the agent share user data with third-party services or vendors?

Yes / NoevidenceTeam+ · GDPR pack

Guardrail impact: Sets privacy.third_party_sharing and marks the audit posture as full_content. Any transfer to external processors must be covered by a DPA.

PV4

Could this agent process data belonging to minors?

Yes / No + follow-upenforcedTeam+ · GDPR pack

Guardrail impact: Sets pii=block, adds FERPA to the framework list and records processes_minor_data. PV4_age_band captures the age band (COPPA under 13, GDPR Art. 8 under 16).

PV7

Has a DPIA (Data Protection Impact Assessment) been conducted for this agent?

Single-select + follow-upevidenceTeam+ · GDPR packOptions: completed, in_progress, not_required, not_applicable

Guardrail impact: Records dpia_status (and PV7_date). DPIAs are mandatory under GDPR Art. 35 for high-risk processing.

Prompt Security

Hardens the agent against prompt injection, output poisoning and system-prompt leakage.

P1

Will this agent accept input from untrusted sources? (User-submitted text, web scraping, email content)

Yes / Noenforced

Guardrail impact: Tunes injection-detection sensitivity — detection always runs. Yes sets injection=strict and llm_classification_required.

P3

Will the agent's system prompt contain sensitive instructions? (API keys, internal URLs, business logic)

Yes / Noenforced

Guardrail impact: Sets prompt_leakage_detection and llm_classification_required — output is monitored for fragments matching the system prompt.

P4_categories

What should this agent never do? (select all that apply)

Multi-selectenforcedOptions: no_delete, no_external_comms, no_pii_storage, no_code_execution, no_credential_access, no_data_export, no_payment_action, no_permission_grants, other

Guardrail impact: Each category compiles to a hard constraint in the guardrail profile and is checked at runtime. Use these structured categories for anything that must be enforced.

P4

Additional constraint notes

Free textadvisory

Guardrail impact: Captured for audit only. This text is not injected into the scanner and is not enforced at runtime.

P5

Will this agent process email or file attachments from untrusted senders?

Yes / Noenforced

Guardrail impact: Raises injection to strict. Attackers embed injection payloads in images and attachments.

P6

Can this agent's output be fed as input to another system or agent without human review?

Yes / No + follow-upenforced

Guardrail impact: Sets output_integrity_check and inter_step_scanning and redacts sensitive output. P6_consumers lists the downstream consumers. Covers multi-step chains and agents whose output triggers automation.

P7

Which prompt-injection encoding defenses should be applied to this agent's inputs?

Multi-selectenforcedOptions: unicode_normalize, strip_zero_width, base64_detect, translation_cross_check, homoglyph_detect

Guardrail impact: Sets scanner encoding_defenses. Each defense catches a specific keyword-filter bypass family.

Security Context

Primes CleverGuard with agent-specific threat intelligence. Free-text answers here are advisory context; the structured fields next to them are what compile into enforced allowlists and detection rules.

S_AUTH

Who is authorized to give this agent instructions?

Free textadvisory

Guardrail impact: Advisory context for authority-spoofing detection (authorized_sources). Complete S_AUTH_structured for the enforced allowlist.

S_AUTH_structured

Authorized instruction sources (structured)

Multi-selectenforcedOptions: internal_employees_only, authenticated_customers, specific_roles, specific_channels, external_partners, anonymous

Guardrail impact: Enforced at the session gate. An instruction claiming to come from a source not on this list is flagged as authority spoofing.

S_FORBID

What data should this agent NEVER include in its output?

Free textadvisory

Guardrail impact: Populates never_output in the security context and defines exfiltration boundaries — output matching these descriptions is flagged or blocked. Cross-referenced with P4_categories.

S_SCOPE

What is this agent's authorized operational scope?

Free textadvisory

Guardrail impact: Populates operational_scope. Actions outside this declared boundary trigger a drift alert.

S_THREATS

What manipulation tactics might target this agent?

Free textadvisory

Guardrail impact: Populates manipulation_tactics — primes the scanner with attack patterns specific to this agent's threat model.

S_THREATS_structured

Expected attack vectors (structured)

Multi-selectenforcedOptions: prompt_injection_direct, prompt_injection_indirect, authority_spoofing, jailbreak_roleplay, encoding_bypass, prompt_leak, data_exfil, insider_threat

Guardrail impact: Each selected vector activates the matching CleverGuard detection rules for this agent.

S_NORMAL

What does a normal request to this agent look like?

Free textadvisory

Guardrail impact: Populates normal_request_pattern — a baseline for anomaly detection. Requests that deviate significantly are flagged for review.

S_VOCAB

What domain-specific terms are normal for this agent?

Free textadvisory

Guardrail impact: Populates domain_vocabulary to prevent false positives on legitimate business vocabulary. Do not paste credentials, internal URLs or sensitive data here.

S_MULTI_AGENT

Does this agent interact with other AI agents or participate in multi-agent orchestration?

Yes / No + follow-upenforced

Guardrail impact: Sets multi_agent_orchestration, inter_agent_scanning and allowed_peers (from S7_agents), and raises injection to strict. Multi-agent workflows introduce confused-deputy attack surfaces.

S_SPAWN

Does this agent spawn other agents or new LLM sessions at runtime?

Yes / Noenforced

Guardrail impact: Sets spawn_allowed. Distinct from T9 (spawn authorization): this captures observed runtime behaviour and is compared against T9 by drift detection.

S_PEER

Which peer agents does this agent communicate with at runtime?

Free textadvisory

Guardrail impact: Used by drift detection to identify undeclared agent-to-agent communication.

Note
The wizard also has steps that are not question groups: Environment (non-prod, production or both), Operations (a free-text description of the agent's expected workload), Triggers (how the agent is activated), Pattern Tuning, Review, Key Access (provider key bindings) and Verify. Operations and Triggers are informational for reviewers and do not directly map to guardrail rules.