/

/

Penetration Testing RFP Template for 2026: Questions, Scorecard and AI Vendor Checklist

Penetration Testing RFP Template for 2026: Questions, Scorecard and AI Vendor Checklist

AI Pentesting

Amartya | CodeAnt AI Code Review Platform

Sonali Sood

Founding GTM, CodeAnt AI

CONTENTS

0% READ
Find Vulnerabilities. Prove They're Exploitable.

Validate real-world vulnerabilities with AI-powered penetration testing.

A penetration testing RFP should make every vendor commit, in writing, to five things: exactly what they will test, how each finding is proven exploitable, what the report gives your auditor, how retests work, and what happens to your data.

In 2026 it needs a sixth: what an AI agent does on its own and who reviews its work. CodeAnt AI fits teams that want exploit-proven findings and a SOC 2 and ISO 27001-ready report within 48 hours; teams that want a named human team on a multi-week engagement may prefer a human-led firm. Below you will find the RFP sections, the questions to ask, a scoring rubric and a side-by-side look at six providers.

This guide gives you a complete RFP structure with specific requirements for evaluating modern pentesting: code-aware methodology, performance-based pricing, and deliverables developers can actually use. You will learn how to scope complex environments, evaluate vendor depth, and structure a contract that aligns with security outcomes rather than page count.

Where this connects: an RFP decides who you hire. The contract clauses decide what they actually test once hired. This guide covers the first half. Pair it with the contract guide for the second.

Pre-RFP Alignment: The Internal Decisions Vendors Cannot Price Around

The most common failure mode is a three-way disconnect: security wants depth, procurement wants a fixed price, and engineering wants minimal disruption. These are not incompatible, but they need explicit decisions before any vendor proposes.

Define your primary objective

Pick one. You cannot optimize for all three at once, and a vendor priced for one will underdeliver on the others.

  • Compliance-focused testing. Goal: audit evidence for SOC 2, ISO 27001, PCI DSS, or HIPAA. Scope: external systems, black box. Timeline: annual or semi-annual. The cost driver is report quality and auditor acceptance.

  • Risk-reduction testing. Goal: find exploitable vulnerabilities before attackers do. Scope: full stack, authenticated workflows, business logic. Timeline: quarterly or continuous. The cost driver is exploitation depth and attack-chain construction.

  • Pre-launch validation. Goal: validate a new feature before release. Scope: a specific feature set, time-boxed. Timeline: sprint-aligned, one-to-two-week turnaround. The cost driver is speed and availability.

Map your technical constraints

Document the operational reality in the RFP so vendors price against facts, not assumptions.

  • Production testing windows: approved hours, blackout periods, rate-limit thresholds, acceptable performance impact, rollback procedures.

  • Environment access: is production testing allowed, does staging mirror production, is test data synthetic or anonymized.

  • Authentication: test accounts per role, MFA handling (bypass or static TOTP seeds), API authentication (OAuth flows, JWT tokens).

A concrete constraint statement removes more ambiguity than a page of prose:

Production testing permitted during maintenance windows (Saturdays 2-6 AM EST)
with 100 req/sec rate limits. Staging available 24/7 but uses t3.medium instances
(vs. prod's c5.2xlarge) and 90-day-old anonymized data. Five test accounts provided
(admin, manager, employee, contractor, guest) with MFA disabled. WAF blocks after
500 requests from single IP in 5 minutes

Production testing permitted during maintenance windows (Saturdays 2-6 AM EST)
with 100 req/sec rate limits. Staging available 24/7 but uses t3.medium instances
(vs. prod's c5.2xlarge) and 90-day-old anonymized data. Five test accounts provided
(admin, manager, employee, contractor, guest) with MFA disabled. WAF blocks after
500 requests from single IP in 5 minutes

Production testing permitted during maintenance windows (Saturdays 2-6 AM EST)
with 100 req/sec rate limits. Staging available 24/7 but uses t3.medium instances
(vs. prod's c5.2xlarge) and 90-day-old anonymized data. Five test accounts provided
(admin, manager, employee, contractor, guest) with MFA disabled. WAF blocks after
500 requests from single IP in 5 minutes

Assign remediation ownership before you issue the RFP

Decide who triages findings, who implements fixes, and what the timeline expectations are (critical in 7 days, high in 30). Settle the retesting logistics too: unlimited retesting or a per-fix charge, turnaround (24 to 48 hours vs a week), and whether you want continuous testing or quarterly re-scans. A finding with no owner and no retest path is a finding that ages into next year's audit.

What Sections Should a Penetration Testing RFP Include?

A complete RFP has ten sections. The two most often missing are the evidence standard (section 6) and data handling and vendor controls (section 9), and they are the two that separate a real test from scanner output in 2026.

  1. Company background and context: what you build, who uses it, and how often you ship.

  2. Objective and success criteria: the framework or customer requirement behind the test, the audit or deadline date, and what a good outcome looks like.

  3. In-scope assets and boundaries: root domains, applications, APIs, cloud accounts, and the user roles and test accounts the tester may use.

  4. Out-of-scope items and restrictions: systems not to touch, production limits and blackout windows.

  5. Testing approach: black box, grey box or white box, defined in words. Black box means no credentials; grey box means credentials for named roles; white box means source code and cloud access.

  6. Evidence and validation standard: every finding must be reproducible from the exact request and response, and the vendor must say who validates findings before you see them.

  7. Deliverables: an executive summary you can share with auditors and customers, a technical report for engineers, and mapping to your framework's current control numbers.

  8. Retest policy: how many retests, within what window, and whether a failed retest costs extra.

  9. Data handling, safety and the vendor's own controls: rules of engagement, how to stop a test, the request log, data retention and deletion, the vendor's own SOC 2 report, tester background checks, and whether any AI model is trained on your data.

  10. Pricing model and contract terms: what the price is based on, what is included, and the clauses listed further down this page.

NIST's guide to security testing (SP 800-115, Appendix B) adds logistics, a communication plan and incident handling to the rules of engagement. Put those in the contract, not the RFP.

Scope: Defining the Modern Attack Surface

Precise scope is what determines whether you get shallow scanning or deep testing. Two worked examples show the level of detail to aim for.

A SaaS web app plus REST API:

Assets:
- Web application: React SPA at app.example.com, 45 routes, 3 user roles
- REST API: 127 endpoints at api.example.com/v2 (/users, /projects, /billing, /admin)
- Authentication: Auth0 OAuth 2.0 with JWT tokens
- Integrations: Stripe payments, SendGrid email, AWS S3 file storage

Testing Mode: Gray box (credentials for all 3 roles, OpenAPI 3.0 spec,
read access to routing logic and auth middleware)

Critical Flows:
- Multi-tenant isolation (can User A access User B's projects?)
- Role-based access (can viewer access admin endpoints?)
- Payment flow (pricing manipulation, payment bypass?)
- File upload (malicious files, unauthorized access?)

Out of Scope: marketing site, internal admin tools, DoS, social engineering
Assets:
- Web application: React SPA at app.example.com, 45 routes, 3 user roles
- REST API: 127 endpoints at api.example.com/v2 (/users, /projects, /billing, /admin)
- Authentication: Auth0 OAuth 2.0 with JWT tokens
- Integrations: Stripe payments, SendGrid email, AWS S3 file storage

Testing Mode: Gray box (credentials for all 3 roles, OpenAPI 3.0 spec,
read access to routing logic and auth middleware)

Critical Flows:
- Multi-tenant isolation (can User A access User B's projects?)
- Role-based access (can viewer access admin endpoints?)
- Payment flow (pricing manipulation, payment bypass?)
- File upload (malicious files, unauthorized access?)

Out of Scope: marketing site, internal admin tools, DoS, social engineering
Assets:
- Web application: React SPA at app.example.com, 45 routes, 3 user roles
- REST API: 127 endpoints at api.example.com/v2 (/users, /projects, /billing, /admin)
- Authentication: Auth0 OAuth 2.0 with JWT tokens
- Integrations: Stripe payments, SendGrid email, AWS S3 file storage

Testing Mode: Gray box (credentials for all 3 roles, OpenAPI 3.0 spec,
read access to routing logic and auth middleware)

Critical Flows:
- Multi-tenant isolation (can User A access User B's projects?)
- Role-based access (can viewer access admin endpoints?)
- Payment flow (pricing manipulation, payment bypass?)
- File upload (malicious files, unauthorized access?)

Out of Scope: marketing site, internal admin tools, DoS, social engineering

Microservices behind an API gateway:

Assets:
- API Gateway: gateway.example.com (Kong)
- Microservices: 52 services (user-service, payment-service, etc.)
- Service mesh: Istio for inter-service communication
- Cloud: AWS (EKS, RDS, S3, Lambda, SQS)

Testing Mode: Gray box with infrastructure access (gateway admin creds,
K8s read access, source for all services, AWS console read-only)

Critical Attack Vectors:
- API Gateway bypass (direct service access?)
- Service-to-service authentication
- Secrets management (environment variables vs. vaults)
- Container escape and lateral movement
- IAM role privilege escalation
Assets:
- API Gateway: gateway.example.com (Kong)
- Microservices: 52 services (user-service, payment-service, etc.)
- Service mesh: Istio for inter-service communication
- Cloud: AWS (EKS, RDS, S3, Lambda, SQS)

Testing Mode: Gray box with infrastructure access (gateway admin creds,
K8s read access, source for all services, AWS console read-only)

Critical Attack Vectors:
- API Gateway bypass (direct service access?)
- Service-to-service authentication
- Secrets management (environment variables vs. vaults)
- Container escape and lateral movement
- IAM role privilege escalation
Assets:
- API Gateway: gateway.example.com (Kong)
- Microservices: 52 services (user-service, payment-service, etc.)
- Service mesh: Istio for inter-service communication
- Cloud: AWS (EKS, RDS, S3, Lambda, SQS)

Testing Mode: Gray box with infrastructure access (gateway admin creds,
K8s read access, source for all services, AWS console read-only)

Critical Attack Vectors:
- API Gateway bypass (direct service access?)
- Service-to-service authentication
- Secrets management (environment variables vs. vaults)
- Container escape and lateral movement
- IAM role privilege escalation

Methodology: Separating Deep Testing From Scanner Output

The methodology section is where you filter out the vendors who will hand you scanner output with a logo on it.

Vendor must describe testing methodology mapped to PTES phases. Explain
specific techniques, tools, and deliverables for each phase. Proposals
that skip threat modeling, exploitation validation, or post-exploitation
will be disqualified

Vendor must describe testing methodology mapped to PTES phases. Explain
specific techniques, tools, and deliverables for each phase. Proposals
that skip threat modeling, exploitation validation, or post-exploitation
will be disqualified

Vendor must describe testing methodology mapped to PTES phases. Explain
specific techniques, tools, and deliverables for each phase. Proposals
that skip threat modeling, exploitation validation, or post-exploitation
will be disqualified

  • Require attack-chain construction. The defining trait of professional pentesting is chaining multiple issues into a working exploit that demonstrates real business impact, not a flat list of isolated findings.

Vendor must demonstrate capability to construct multi-step attack chains.
Provide examples of past findings where you chained 2+ vulnerabilities
(e.g., IDOR + privilege escalation, SSRF + credential exposure) to achieve
unauthorized access or data exfiltration

Vendor must demonstrate capability to construct multi-step attack chains.
Provide examples of past findings where you chained 2+ vulnerabilities
(e.g., IDOR + privilege escalation, SSRF + credential exposure) to achieve
unauthorized access or data exfiltration

Vendor must demonstrate capability to construct multi-step attack chains.
Provide examples of past findings where you chained 2+ vulnerabilities
(e.g., IDOR + privilege escalation, SSRF + credential exposure) to achieve
unauthorized access or data exfiltration

  • Require business logic and authorization testing. These are exactly the flaws automated scanners cannot reason about, so make them explicit RFP questions: how the vendor tests business logic (price manipulation, workflow bypass, race conditions), how they test Broken Object Level Authorization and IDOR across APIs, and an example of an authorization flaw they found that required understanding the application's data model and role structure.

  • Require a code-aware explanation. If you are giving source access, make the vendor earn it by explaining how they use it: tracing data flows, identifying authentication and authorization middleware, discovering undocumented endpoints, and understanding business logic implementation, with examples of vulnerabilities found through code analysis that external-only testing would have missed.

Deliverables: What an Audit-Grade Report Actually Contains

Split the report for two audiences. The executive summary carries business impact, compliance-control violations, aggregate risk, and a remediation timeline. The engineering appendix carries file-and-line code references, reproducible PoC exploits, CVSS scores with justification, step-by-step remediation with code, and verification steps.

Per finding, insist on all five of these. A reproducible PoC:

# Broken Object Level Authorization
curl -X GET 'https://api.example.com/users/742/profile' \
  -H 'Authorization: Bearer <user_123_token>'
# Expected: 200 OK with user 742's data (should be 403 Forbidden)
# Broken Object Level Authorization
curl -X GET 'https://api.example.com/users/742/profile' \
  -H 'Authorization: Bearer <user_123_token>'
# Expected: 200 OK with user 742's data (should be 403 Forbidden)
# Broken Object Level Authorization
curl -X GET 'https://api.example.com/users/742/profile' \
  -H 'Authorization: Bearer <user_123_token>'
# Expected: 200 OK with user 742's data (should be 403 Forbidden)

A CVSS v3.1 score with its vector, so the severity is auditable rather than asserted:

CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:C/C:H/I:H/A:N  (9.6, Critical)
CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:C/C:H/I:H/A:N  (9.6, Critical)
CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:C/C:H/I:H/A:N  (9.6, Critical)

Code-level remediation that shows the fix, not just the flaw:

# Before (vulnerable)
def get_profile(user_id):
    return User.query.get(user_id)

# After (secure)
def get_profile(user_id):
    if current_user.id != user_id and not current_user.is_admin:
        abort(403)
    return User.query.get(user_id)
# Before (vulnerable)
def get_profile(user_id):
    return User.query.get(user_id)

# After (secure)
def get_profile(user_id):
    if current_user.id != user_id and not current_user.is_admin:
        abort(403)
    return User.query.get(user_id)
# Before (vulnerable)
def get_profile(user_id):
    return User.query.get(user_id)

# After (secure)
def get_profile(user_id):
    if current_user.id != user_id and not current_user.is_admin:
        abort(403)
    return User.query.get(user_id)

And a verification step so your team can confirm the fix closed the issue:

curl -X GET 'https://api.example.com/users/742/profile' \
  -H 'Authorization: Bearer <user_123_token>'
# Expected: 403 Forbidden
curl -X GET 'https://api.example.com/users/742/profile' \
  -H 'Authorization: Bearer <user_123_token>'
# Expected: 403 Forbidden
curl -X GET 'https://api.example.com/users/742/profile' \
  -H 'Authorization: Bearer <user_123_token>'
# Expected: 403 Forbidden

Finally, require compliance mapping in the report itself, since assembling it afterward is what turns a clean test into a scramble before the audit.

Framework

What it says about testing

What to ask for in the RFP

SOC 2 (AICPA Trust Services Criteria)

CC4.1: the entity selects, develops and performs ongoing and/or separate evaluations. CC7.1 covers detecting new vulnerabilities. SOC 2 does not require a pentest; a pentest is one way to support these criteria.

A report that states scope, dates, methodology and remediation status, plus an executive summary you can hand to the auditor.

ISO/IEC 27001:2022

Annex A 8.8 (technical vulnerability management) and Annex A 8.29 (security testing in development and acceptance).

Findings mapped to the 2022 Annex A numbers your certificate uses, not the 2013 numbering (A.12.6.1, A.14.2.8).

PCI DSS v4.0.1

11.4.1: a defined, industry-accepted methodology. 11.4.2 and 11.4.3: internal and external tests at least once every 12 months, by a qualified internal resource or qualified external third party, with organizational independence of the tester. 11.4.4: exploitable findings corrected and testing repeated to verify. Quarterly scans by an Approved Scanning Vendor are a separate requirement (11.3.2).

The tester's qualifications, a statement of organizational independence, coverage of the cardholder data environment perimeter and segmentation controls, and a retest that verifies corrections.

HIPAA Security Rule

45 CFR 164.308(a)(1)(ii)(A): an accurate and thorough risk analysis. 164.308(a)(8): a periodic technical and nontechnical evaluation. The rule does not name penetration testing.

A report your risk analysis can cite: systems tested, exploitable risks to systems that hold ePHI, and remediation status.

2026-Specific Requirements: The Modern Attack Surface

A generic RFP template written five years ago misses most of what actually gets breached today. Add explicit requirements for four modern surfaces.

  • API and microservices. Require GraphQL testing for Broken Object Level Authorization across every query and mutation, including query manipulation, field injection, and batching attacks, with the specific queries that retrieved unauthorized data. Require gRPC testing that analyzes .proto definitions to enumerate RPC methods and tests authorization on each. And require service-to-service testing that attempts to forge service identity tokens and bypass service-mesh policies to show lateral movement. The cloud pentesting guide and the hybrid attack-chain guide cover these paths in depth.

  • Cloud identity and infrastructure. Require three things here: secrets scanning across repositories, container images, CI/CD pipelines, and cloud storage, cloud IAM testing that attempts privilege escalation from an application service account to admin, and CI/CD testing that reports the blast radius of a compromised pipeline.

  • AI and LLM security. This is the newest gap in most templates. Require prompt-injection testing (both direct user input and indirect poisoned data sources), data-leakage testing for PII and secrets in model responses, and tool-abuse testing for LLMs with function-calling, showing how tool abuse chains into broader compromise. Be honest with yourself about scope here: many traditional firms have no LLM-testing capability at all, so if you run AI features, make it a named requirement rather than an assumption.

What Should You Ask an AI or Autonomous Pentest Vendor?

Most pentest RFP templates were written for human testers. When an AI agent does the testing, you also need to know what it does on its own, how it is kept in scope, and what record it leaves. The OWASP Autonomous Penetration Testing Standard (APTS), an OWASP incubator project at version 0.1.0, groups governance requirements for these platforms into eight domains. Use them as the spine of the AI section of your RFP.

APTS domain

Question to put in the RFP

A good answer includes

Scope enforcement

How does the agent stay inside the authorized scope, and what happens when it finds an asset outside it?

A technical control, such as an allowlist of hosts, and out-of-scope assets reported to you rather than tested.

Safety controls

Which actions are off by default, and how do we stop a test at once?

Named blocked actions, a stop method you can trigger, and who answers during the test.

Human oversight

Which findings does a person review before they reach us, and who signs the report?

A defined review step and the reviewer's role named on the report.

Graduated autonomy

What runs without approval, and can we start with less autonomy?

Settings you can choose, for example a black box pass without credentials before any grey box testing.

Auditability

Can we get the log of every request, response and decision, and how long is it kept?

An exportable log tied to each finding, with the retention period written into the contract.

Manipulation resistance

How does the agent handle instructions planted in our own pages or API responses?

A described control against prompt injection aimed at the agent itself.

Supply chain trust

Which AI models and subprocessors see our data, and is any of it used for training?

A subprocessor list, the model providers used, and a no-training clause in the contract.

Reporting

Show us one real finding with the exact request that reproduces it.

A redacted sample finding with the request, the response, the impact and the fix.

Also ask whether the vendor has assessed its platform against the APTS checklists, and at which of the standard's three tiers.

The auditability question matters more than it looks. In sales conversations this autumn, a fintech buyer comparing AI pentest vendors ruled one out because its deliverable read like a generated document and the vendor could not say where the test records were kept. A log of what was tested, and where testing stopped, is what turns an AI pentest into evidence.

How Should You Score Pentest RFP Responses?

Score evidence first. A vendor that cannot show how a finding was proven, and what it tested, cannot be compared on anything else.

Criterion

Weight

What earns full marks

Evidence and validation

30%

Every finding reproducible from a request and response; a named review step; a sample report with real, redacted findings

Scope fit and testing depth

20%

Covers your apps, APIs, cloud and user roles; black, grey or white box defined in writing; findings chained into attack paths

Safety, data handling and vendor controls

15%

Rules of engagement, a stop method, a request log, retention and deletion terms, the vendor's own SOC 2 report and tester vetting

Report fit for auditors and customers

15%

A shareable executive summary; mapping to your framework's current control numbers; scope, dates and methodology stated

Retests and remediation

10%

Retest count and window in writing; fix guidance engineers can act on

Commercial terms

10%

Price tied to your scope; what is and is not included; no surprises on unused credits or extra retests

Disqualify a response when the vendor accepts a vague scope without questions, will not share a sample report, describes its methodology only as proprietary, cannot say where test records are kept, or leaves retest terms out.

Weight a response up when the vendor asks about your roles, tenants and login flows during scoping, shows the reproducing request for each sample finding, puts retests in writing for a stated period, and shares its own SOC 2 report and data deletion process.

How Do You Run a Fair Side-by-Side Test Before You Sign?

Buyers we speak with increasingly shortlist two vendors and test both before committing. The comparison is only fair if both vendors get the same system, the same parameters, the same credentials and the same window. Then compare:

  • Exploitable findings with evidence, not the total number of findings.

  • Attack paths that reach real data, such as a low-privilege user becoming an admin.

  • False positives you had to throw away.

  • Time from kickoff to report, and time from fix to verified retest.

  • Whether the report is something your auditor and your customers can read.

Three patterns from those conversations are worth planning for. Teams under a board or investor mandate need to show they compared vendors, so document the comparison as you go. Teams with a long-standing incumbent often move one or two tests to the new vendor and compare results after the incumbent's next test. And some enterprises rotate vendors per application every year, so the RFP gets reused annually; keep it as a living document. Start the comparison at least a month before your next scheduled test so the audit date does not force the decision.

How do Six Pentest Providers Answer the Core RFP Questions?

Here is how six providers describe their answers to the questions above on their own public pages, checked on November 11, 2026. Use it to pre-fill your shortlist, then make each vendor confirm in writing.

Provider

Model

Time to report

Pricing model

Retests

Report and compliance mapping

Main tradeoff

CodeAnt AI

AI agents attack like an adversary and chain steps; a researcher checks every critical finding; black, grey and white box

Within 48 hours

Priced on the attack surface you have, with very generous pricing for fast-growing companies

Unlimited retests for 12 months

SOC 2 and ISO 27001-ready report; every request, response and decision point logged

No public price list; you get a quote scoped to your attack surface

Cobalt

PTaaS with human testers, plus an autonomous pentest overseen by Cobalt testers

Autonomous pentest findings in 24 hours; tiered tests start in 1 to 3 business days

Credits in annual packages (1 credit = 8 hours of testing); autonomous pentest $3,500 per test, limited offer to Dec 31, 2026

Unlimited, on demand for the contract term

Customizable reports; frameworks not named on the pricing page

Credits do not roll over and are spent on hours, not on findings

XBOW

Autonomous; no human drives the attack; findings reviewed before you see them

First validated findings in hours; no report deadline stated

No price on xbow.com; the main call to action is a demo

Free and unlimited

Lists SOC 2, ISO 27001, ISO 42001, PCI DSS and NIS 2 under deployment and compliance

No public price and no stated report deadline

Aikido

Fixed-scope Typical pentest, or a Rightsized pentest scoped to your application

Same-day results

Typical pentest $4,000 for one application and its primary APIs; "No High or Critical Finding = Don't Pay"

Free for up to 6 months

Full PDF report for SOC 2 and ISO 27001

Typical package is fixed to one application; human review not described on the pricing page

Synack

Sara: AI-led; SynackST: one human tester; Synack14: a team of testers

4 to 5 business day assessment window (Sara); 5 day window (SynackST)

Sara from $5,085, SynackST from $10,283, Synack14 from $27,120; platform billed separately

Patch verification on all tiers; no count stated

Compliance-ready report on SynackST and above; Sara gets a comprehensive report

AI tier's report is not labeled compliance-ready; platform is an extra line item

Astra

Pentest Auto: autonomous; Pentest Expert: certified testers plus agents

Same day (Auto); manual pentests 10 to 15 working days

Auto $2,999 a year; Expert $5,999 a year; Enterprise from $9,999 a year

1 to 4 re-scans in a 30 or 90 day window, by plan

SOC 2, ISO 27001 and HIPAA reports; CREST, PCI ASV and CERT-In reports on Expert and above

Short re-scan windows; manual tests take two to three weeks

Which pricing model should your RFP ask for, and which clauses protect you?

Pentest pricing in 2026 falls into four models. Ask every vendor to state which one they use and what it means for a test that finds little.

  • Credits or hours. Cobalt sells credits in annual packages, where one credit equals 8 hours of testing and credits do not roll over. Ask what happens to unused credits and what a finding-free test costs.

  • Per test, fixed scope. Synack's Sara pentest starts at $5,085 per test and Astra's Pentest Expert at $5,999 a year (both checked November 11, 2026). Ask exactly which apps, APIs and roles the fixed scope covers.

  • Outcome-linked. Aikido's typical pentest is $4,000 for one application and its primary APIs, with "No High or Critical Finding = Don't Pay". Ask who decides severity and how high and critical are defined.

  • Time and materials or retainer. Useful for exploratory work or frequent releases. Ask for a cap and a weekly summary of hours.

CodeAnt prices each pentest on the attack surface you have, and gives very generous pricing to fast-growing companies.

Put these clauses in the contract, not just the proposal:

  1. Scope change: how new assets are added and priced mid-test.

  2. Report date: the committed delivery date for the report, and for critical findings flagged during the test.

  3. Retests: count, window and cost of a failed retest.

  4. Stop procedure: who can halt testing and how fast it stops.

  5. Data: where test data is stored, how long it is kept, proof of deletion, and a ban on training AI models with your data.

  6. Confidentiality and liability: a nondisclosure agreement and responsibility for damage caused by testing. NIST SP 800-115 (section 6.4.1) notes that external assessors may be responsible for damages and should sign any required nondisclosure agreements.

  7. Report signer: the firm and the role that signs the report your auditor or customer will read.

A Worked Example: A Filled-In Mini-RFP

Every section above is a rule. Here is what they look like assembled into one short, sendable RFP for a fictional SaaS company, Northwind Health, a HIPAA-regulated telehealth platform. Lift the structure, swap in your own specifics.

PENETRATION TESTING RFP, Northwind Health Inc.

1. Context
Northwind Health operates a HIPAA-regulated telehealth SaaS handling PHI for
~120,000 patients. We deploy 8-12 times per week across a React front end and a
microservices back end. We are issuing this RFP for our annual HIPAA/SOC 2
assessment plus risk-reduction testing of our new provider-messaging API.

2. Primary objective
Risk reduction (primary) with compliance evidence (secondary). We want exploitable
findings on the messaging API, plus SOC 2 CC6.1/CC6.6 and HIPAA 164.312(a)(1)
control mapping in the report.

3. In-scope assets
- Web app: app.northwind.health (React SPA, 52 routes, 4 roles:
  patient, provider, admin, billing)
- REST API: api.northwind.health/v3 (94 endpoints; /patients, /messages,
  /appointments, /billing, /admin)
- New: provider-messaging service (GraphQL, 18 queries, 11 mutations)
- Auth: Okta OIDC + JWT
- Cloud: AWS (EKS, RDS Postgres, S3 for attachments)
- Integrations: Stripe, Twilio, a third-party e-prescribing API

4. Out of scope
Marketing site, the Twilio and e-prescribing third-party platforms themselves,
DoS testing, social engineering, physical security.

5. Testing mode
Code-aware gray box. We provide credentials for all 4 roles, the OpenAPI 3.0
and GraphQL schemas, read access to the messaging service source, and read-only
AWS console access.

6. Critical flows (must be tested)
- Cross-tenant PHI access (can provider A read patient B's messages?)
- Role escalation (billing role reaching /admin?)
- GraphQL BOLA across all 18 queries and 11 mutations
- S3 attachment authorization (can a patient fetch another patient's file?)
- Payment manipulation on the Stripe flow

7. Deliverables (per finding)
Working PoC (curl/script), CVSS v3.1 vector + score, file/line remediation,
business impact, and mapping to SOC 2 CC6.1/CC6.6 and HIPAA 164.312(a)(1).
Findings without a working PoC are classified Informational and do not count
toward the engagement fee.

8. Retest and pricing
Unlimited retesting for 90 days, 48-hour turnaround, no per-retest fee.
Performance-based pricing preferred: we pay for confirmed exploitable
high/critical findings with a working PoC.

9. Constraints
Production testing Saturdays 2-6 AM ET, 100 req/sec cap, WAF blocks at 500
req/5 min/IP. Staging available 24/7 with anonymized data. MFA disabled on
the five provided test accounts.

10. Evaluation
Weighted scorecard: technical capability 35%, team expertise 25%, cost/value
20%, methodology depth 10%, cultural fit 10%. A sample report and a walkthrough
of one past multi-step attack chain are required before shortlisting

PENETRATION TESTING RFP, Northwind Health Inc.

1. Context
Northwind Health operates a HIPAA-regulated telehealth SaaS handling PHI for
~120,000 patients. We deploy 8-12 times per week across a React front end and a
microservices back end. We are issuing this RFP for our annual HIPAA/SOC 2
assessment plus risk-reduction testing of our new provider-messaging API.

2. Primary objective
Risk reduction (primary) with compliance evidence (secondary). We want exploitable
findings on the messaging API, plus SOC 2 CC6.1/CC6.6 and HIPAA 164.312(a)(1)
control mapping in the report.

3. In-scope assets
- Web app: app.northwind.health (React SPA, 52 routes, 4 roles:
  patient, provider, admin, billing)
- REST API: api.northwind.health/v3 (94 endpoints; /patients, /messages,
  /appointments, /billing, /admin)
- New: provider-messaging service (GraphQL, 18 queries, 11 mutations)
- Auth: Okta OIDC + JWT
- Cloud: AWS (EKS, RDS Postgres, S3 for attachments)
- Integrations: Stripe, Twilio, a third-party e-prescribing API

4. Out of scope
Marketing site, the Twilio and e-prescribing third-party platforms themselves,
DoS testing, social engineering, physical security.

5. Testing mode
Code-aware gray box. We provide credentials for all 4 roles, the OpenAPI 3.0
and GraphQL schemas, read access to the messaging service source, and read-only
AWS console access.

6. Critical flows (must be tested)
- Cross-tenant PHI access (can provider A read patient B's messages?)
- Role escalation (billing role reaching /admin?)
- GraphQL BOLA across all 18 queries and 11 mutations
- S3 attachment authorization (can a patient fetch another patient's file?)
- Payment manipulation on the Stripe flow

7. Deliverables (per finding)
Working PoC (curl/script), CVSS v3.1 vector + score, file/line remediation,
business impact, and mapping to SOC 2 CC6.1/CC6.6 and HIPAA 164.312(a)(1).
Findings without a working PoC are classified Informational and do not count
toward the engagement fee.

8. Retest and pricing
Unlimited retesting for 90 days, 48-hour turnaround, no per-retest fee.
Performance-based pricing preferred: we pay for confirmed exploitable
high/critical findings with a working PoC.

9. Constraints
Production testing Saturdays 2-6 AM ET, 100 req/sec cap, WAF blocks at 500
req/5 min/IP. Staging available 24/7 with anonymized data. MFA disabled on
the five provided test accounts.

10. Evaluation
Weighted scorecard: technical capability 35%, team expertise 25%, cost/value
20%, methodology depth 10%, cultural fit 10%. A sample report and a walkthrough
of one past multi-step attack chain are required before shortlisting

PENETRATION TESTING RFP, Northwind Health Inc.

1. Context
Northwind Health operates a HIPAA-regulated telehealth SaaS handling PHI for
~120,000 patients. We deploy 8-12 times per week across a React front end and a
microservices back end. We are issuing this RFP for our annual HIPAA/SOC 2
assessment plus risk-reduction testing of our new provider-messaging API.

2. Primary objective
Risk reduction (primary) with compliance evidence (secondary). We want exploitable
findings on the messaging API, plus SOC 2 CC6.1/CC6.6 and HIPAA 164.312(a)(1)
control mapping in the report.

3. In-scope assets
- Web app: app.northwind.health (React SPA, 52 routes, 4 roles:
  patient, provider, admin, billing)
- REST API: api.northwind.health/v3 (94 endpoints; /patients, /messages,
  /appointments, /billing, /admin)
- New: provider-messaging service (GraphQL, 18 queries, 11 mutations)
- Auth: Okta OIDC + JWT
- Cloud: AWS (EKS, RDS Postgres, S3 for attachments)
- Integrations: Stripe, Twilio, a third-party e-prescribing API

4. Out of scope
Marketing site, the Twilio and e-prescribing third-party platforms themselves,
DoS testing, social engineering, physical security.

5. Testing mode
Code-aware gray box. We provide credentials for all 4 roles, the OpenAPI 3.0
and GraphQL schemas, read access to the messaging service source, and read-only
AWS console access.

6. Critical flows (must be tested)
- Cross-tenant PHI access (can provider A read patient B's messages?)
- Role escalation (billing role reaching /admin?)
- GraphQL BOLA across all 18 queries and 11 mutations
- S3 attachment authorization (can a patient fetch another patient's file?)
- Payment manipulation on the Stripe flow

7. Deliverables (per finding)
Working PoC (curl/script), CVSS v3.1 vector + score, file/line remediation,
business impact, and mapping to SOC 2 CC6.1/CC6.6 and HIPAA 164.312(a)(1).
Findings without a working PoC are classified Informational and do not count
toward the engagement fee.

8. Retest and pricing
Unlimited retesting for 90 days, 48-hour turnaround, no per-retest fee.
Performance-based pricing preferred: we pay for confirmed exploitable
high/critical findings with a working PoC.

9. Constraints
Production testing Saturdays 2-6 AM ET, 100 req/sec cap, WAF blocks at 500
req/5 min/IP. Staging available 24/7 with anonymized data. MFA disabled on
the five provided test accounts.

10. Evaluation
Weighted scorecard: technical capability 35%, team expertise 25%, cost/value
20%, methodology depth 10%, cultural fit 10%. A sample report and a walkthrough
of one past multi-step attack chain are required before shortlisting

Notice what this does that a generic template does not: it names the objective and its priority order, enumerates the real surface down to endpoint and role counts, calls out the exact critical flows, ties deliverables to the specific compliance controls the auditor will ask for, and makes the working PoC a payment condition rather than a hope. A vendor who cannot answer this cleanly is a vendor who was going to send you scanner output.

How does CodeAnt AI answer this RFP?

CodeAnt AI runs AI-led penetration tests with black box, grey box and white box scopes. Its agents attack your apps like an adversary would, chain steps and report only what they prove, and a researcher checks every critical finding before it reaches you. Every request, every response and every decision point is logged in sequence, so you can see what was tested and where testing stopped. The report is SOC 2 and ISO 27001-ready within 48 hours, and you get unlimited retests for 12 months after each fix. CodeAnt is SOC 2 Type II and a Gartner Cool Vendor 2026.

CodeAnt AI has run 12,000 pentests and is trusted by more than 1,000 companies, including Fortune 500 companies like NRG, cybersecurity leaders like Commvault, and fast-growing companies like 11x and Sanas.

CodeAnt prices each pentest on the attack surface you have, and gives very generous pricing to fast-growing companies. The tradeoff: there is no public price list, so you get a quote scoped to your attack surface.

Stop Buying Page Count. Buy Proof.

A good RFP is a filter. Written well, it attracts vendors who can test modern architecture and deliver findings your engineers can act on, and it screens out the ones who will send a 200-page scanner dump. Before you send it, make sure four things are non-negotiable: scope precision (enumerate the real surface, not "the application"), evidence-based deliverables (PoC exploits, CVSS with vectors, file-and-line remediation), methodology depth (attack chains, code-aware testing, unlimited retesting), and the 2026 requirements (API, microservices, cloud, and AI/LLM).

For teams deploying weekly or faster, there is a deeper point the RFP process surfaces: an annual engagement is a point-in-time answer to a continuously changing question. That is where CodeAnt AI fits. The same code intelligence that reviews your pull requests conducts the adversarial testing, so it tests from the outside with inside knowledge of your codebase, traces data flows, maps authentication middleware, and builds the attack chains external-only testing misses. Every finding lands with a working PoC and a file-and-line fix, retests are unlimited, and pricing is performance-based, you pay only when a high or critical is confirmed exploitable.

Where to start this week

Take your last pentest RFP (or the generic template you were about to send) and add three requirements to it: a mandatory working PoC for every critical and high finding, a required sample report before shortlisting, and an explicit AI/LLM testing clause if you ship AI features. Those three additions alone will screen out the scanner-only vendors before they ever reach your shortlist.

See how CodeAnt runs a 48-hour pentest →

Related reading

Sources

Frequently Asked
Questions
What should a penetration testing RFP include?

Ten sections: company background, objective and success criteria, in-scope assets, out-of-scope items, the testing approach (black, grey or white box), the evidence standard for each finding, deliverables, retest policy, data handling and vendor controls, and pricing and contract terms. NIST SP 800-115's rules of engagement template adds logistics, communication and incident handling, which belong in the contract.

What should I ask an AI or autonomous pentest vendor?
Is retesting included, or do I pay per retest?
How do I compare two pentest vendors fairly?
What if the pentest finds nothing? Did they miss something?
Go Beyond Detection.
Prove What's Exploitable.