How to read Disclaimer

Part A Β· Foundations

Chapter 2 Β· Law, ethics and rules of engagement

Technical intensity: Low Free preview

Learning outcomes

  1. You can explain why written authorisation is the one fact that separates legal security testing from crime. You can say what happens to testers who lack it, or whose paperwork has gaps.
  2. You can describe India's legal frame at working depth: the Information Technology Act, 2000 (sections 43, 66, 66B–66F, 72A and related), the Digital Personal Data Protection Act, 2023 with its 2025 Rules, and the CERT-In Directions of 2022.
  3. You can outline the frameworks that bind banks and markets β€” RBI's cyber-security framework and SEBI's consolidated CSCRF β€” and the main foreign laws: the US CFAA, the UK Computer Misuse Act, the EU's GDPR and NIS2.
  4. You can assemble the paperwork of a professional engagement: engagement letter, scope, rules of engagement, emergency contacts, data-handling terms. You can apply the stop-and-ask protocol when scope is unclear.
  5. You can take part safely and usefully in bug bounty and coordinated vulnerability disclosure: read a policy's safe harbour, respect disclosure timelines, and report through proper channels including CERT-In.

Concepts

Written permission is everything

Every legal system in this chapter draws the same line, in its own words: access to a computer system is lawful when it is authorised, and unlawful when it is not. Permission separates a penetration test from a break-in, and a security researcher from an intruder. And because memory can be argued with and intentions are invisible, permission that is not written is permission you cannot prove when it matters β€” at 9 p.m. to a deputy, or two years later to a judge.

Written authorisation has three properties. All three must be true. It is informed: the person granting it understands what will be tested, how, and when. It is competent: it comes from someone who owns or controls the systems β€” not a friend in IT, and not a helpful manager two levels below the system owner. It is specific: it names targets, time windows and permitted techniques. Everything outside those boundaries stays what it was before: off-limits. If any property is missing, you do not have authorisation. You have a misunderstanding waiting to become a case.

The Coalfire story adds the practical half of the rule: authorisation must also reach everyone who might enforce the law against you. The contract existed. The sheriff did not know. Professional practice solves this on purpose. The client informs local law enforcement before any physical test. Guards and building staff are briefed. Testers carry a letter with a 24-hour phone number that a doubtful officer can call. You will practise writing exactly that letter in this chapter's core lab.

The Indian frame: Information Technology Act, 2000

India's main cyber law is the Information Technology Act, 2000 (IT Act), heavily amended in 2008. It is the law you must know best. It governs your lab, your future engagements and β€” if things go wrong β€” your freedom. The Act works in two layers for unauthorised access. A civil layer compensates victims. A criminal layer punishes offenders.[3]

Section 43 β€” the civil layer. Anyone who, without permission, accesses a computer, computer system or network β€” or downloads, copies or extracts data, introduces a virus or other contaminant, damages the system, disrupts it, or helps any of this β€” must pay damages as compensation to the affected person. The 2008 amendment removed the earlier money cap, so compensation has no fixed ceiling. This section is why "I only looked around" still costs money. Access without permission is itself the wrong.

Section 66 β€” the criminal layer. When the acts in section 43 are done dishonestly or fraudulently, they become computer-related offences: prison up to three years, or a fine up to β‚Ή5 lakh, or both. "Dishonestly" and "fraudulently" keep their general penal-code meanings: acting to cause wrongful gain or wrongful loss. For you, the practical reading is simple. Unauthorised access done for any advantage, or causing any harm, lands here.

Sections 66B–66F β€” the named offences. Section 66B punishes dishonestly receiving a stolen computer resource or communication device: up to three years, or β‚Ή1 lakh, or both. Section 66C is identity theft β€” fraudulent or dishonest use of someone's electronic signature, password or other unique identification: up to three years and a fine up to β‚Ή1 lakh, the fine mandatory alongside prison. Section 66D is cheating by personation using a computer resource: up to three years and a fine up to β‚Ή1 lakh. Section 66E punishes violation of privacy β€” capturing, publishing or transmitting the image of a person's private area without consent: up to three years, or β‚Ή2 lakh, or both. Section 66F defines cyber terrorism β€” acts that threaten India's sovereignty, integrity or security, or spread terror by denying computer services or causing deaths or damage. Its punishment is imprisonment for life.[3]

Sections 72 and 72A β€” secrets and contracts. Section 72 punishes breach of confidentiality and privacy by a person who had lawful access to records or information under a contract and discloses them without consent: up to two years, or β‚Ή1 lakh, or both. Section 72A extends this to personal information given under a lawful contract for a limited purpose and then disclosed maliciously or in breach of contract: up to three years, or β‚Ή5 lakh, or both. For a working tester these sections are personal. You will hold credentials, extracts and personal data under contract. These sections price a careless leak.

Section 70 β€” protected systems. The government may name certain computer systems "protected systems". Unauthorised access or attempts against them bring prison up to ten years and a fine. The practical lesson: some Indian infrastructure carries much higher stakes, and the exact list is notified officially. No engagement makes such systems fair game without explicit, specific sanction.

A lesson in how cyber law evolves. Section 66A once punished sending "offensive messages". The Supreme Court struck it down in Shreya Singhal v. Union of India (2015) as an unconstitutional limit on free speech. Vague provisions criminalising "annoyance" or "inconvenience" could not survive Article 19(1)(a) of the Constitution.[4] Keep that case in mind as a habit. Statutes are written, challenged and reshaped. The version you learned in college may not be the version in force. Working professionals check the current text on India Code and read the judgments, not summaries.

Unverified: check source The IT Act's section 43A once required "reasonable security practices" for sensitive personal data handled by companies. It was partly superseded as the DPDP regime (below) came into force. Check the current status of any 43A-based duty on India Code or with a lawyer before relying on it.

Protecting people's data: DPDP Act, 2023 and the Rules, 2025

The Digital Personal Data Protection Act, 2023 (DPDP Act) is India's comprehensive data-protection law. Parliament passed it in August 2023. It was brought into force together with the Digital Personal Data Protection Rules in mid-November 2025, with the Rules' duties starting in phases on staggered deadlines.[5][6] The Act covers "digital personal data" β€” data by which a person can be identified. It changes your professional life in three ways.

Consent and notice. Organisations must give a clear notice: what is collected and why. They must obtain consent that is free, specific, informed, unconditional and unambiguous, with the right to withdraw it. The Act allows narrow "legitimate uses", such as responding to a medical emergency or obeying law. For a security professional, breach dumps and incident artefacts are personal data too. The lazy habit of copying "a few sample records" into a report now has a statutory shadow.

Breach notification. When a data breach happens, the data fiduciary must inform the Data Protection Board of India and the affected individuals without delay, describing the facts, likely consequences and the fix. The Rules phase in the exact mechanics of that intimation.[6]

Penalties with teeth. The Data Protection Board administers the penalty schedule. It runs to β‚Ή250 crore for failure to take reasonable security safeguards. Separate, similarly sized slabs cover other failures, such as not meeting children's-data duties or the duties of Significant Data Fiduciaries (which include appointing an India-based data-protection officer and audits).[5] Unverified: check source The exact per-offence slabs and the phased start dates of specific Rules should be checked against the notified text on MeitY's portal before you rely on them professionally. They were newly in force when this book was written.

This is also why this book's examples and labs use only synthetic, clearly fictional personal data. Reproducing real leaked records would itself be a data-protection failure, plus a copyright and ethics problem. Make that discipline yours from day one.

The reporting heartbeat: CERT-In Directions, 2022

On 28 April 2022, the Indian Computer Emergency Response Team (CERT-In) β€” the national incident-response agency set up under section 70B of the IT Act β€” issued binding Directions on information-security practices and cyber-incident reporting. They took effect sixty days later.[7] Two numbers from those Directions matter to every professional in India.

Six hours. The listed incident types β€” which include ransomware, data breaches and leakage, unauthorised access and social-engineering attacks, among others β€” must be reported to CERT-In within six hours of noticing the incident. This applies to service providers, intermediaries, data centres, companies and government organisations.[7][8] Six hours is strict by world standards. It reshapes Indian incident-response plans: detection, triage and the decision to report must all fit inside one evening.

180 days. Covered entities must keep logs of their ICT systems for 180 days within India. ICT stands for Information and Communications Technology. This retention window exists so investigators can rebuild attacks. Chapter 32 returns to it when discussing what a SOC actually keeps.[7][8] Breaking Directions issued under section 70B attracts the Act's residuary penalty (section 45): a fine up to β‚Ή1 lakh, plus a daily amount for continuing default.[3]

Sector regulators: RBI and SEBI

Banks and markets answer to two more masters. The Reserve Bank of India's June 2016 Cyber Security Framework for Banks requires a board-approved cyber-security policy, a cyber-security operations centre, prompt reporting of unusual incidents, and an inventory that maps the bank's assets to risks. This framework turned Indian banking into the country's largest employer of security staff.[9] SEBI's consolidated Cybersecurity and Cyber Resilience Framework (CSCRF), issued in 2024, replaced six older circulars with one graded framework for all its regulated entities. Its phased deadlines were extended by two months to 31 August 2025.[10][11] If your career touches fintech β€” and this book hopes it will β€” these two frameworks are your clients' daily reality. Your engagements will cite them, your findings will map to them, and auditors who check them will read your reports.

The global frame, briefly

Security work crosses borders by default. Your target's cloud sits in one country, your keyboard in another, the attacker in a third. Four foreign instruments matter most in practice.

United States β€” the Computer Fraud and Abuse Act (CFAA), 18 U.S.C. Β§1030. This statute criminalised unauthorised access to "protected computers" β€” a definition broad enough to cover nearly everything connected to the internet. Its most important modern moment is Van Buren v. United States (2021). The Supreme Court held 6–3 that a person entitled to retrieve information from a system does not "exceed authorised access" merely by having an improper purpose for looking at it. The reading is "gates-up-or-down": either you may open the gate or you may not, and your motive does not change the gate.[12][13] For our field the lesson survives the detail: authorisation defines legality, and systems you were never entitled to enter remain criminal to enter. Separately, the US Department of Justice's 2022 charging policy tells prosecutors not to pursue good-faith security research that stays within authorisation. That is a welcome recognition, not a licence.

United Kingdom β€” Computer Misuse Act 1990. The original Western anti-hacking statute. Section 1 criminalises unauthorised access. Section 3 criminalises unauthorised acts meant to impair a system's operation. It reads fresh after three decades, and it applies to actions affecting UK systems from anywhere.

European Union β€” GDPR and NIS2. The General Data Protection Regulation (2016/679) governs personal data, including breach notification to the supervisory authority within 72 hours where risk exists. The NIS2 Directive (2022/2555) imposes cybersecurity duties on essential and important entities, with an early warning within 24 hours and a fuller notification within 72. Multinational engagements often need to respect several of these at once.

Cross-border reality. None of these statutes confines itself to nationals. Accessing an Indian system from abroad breaches the IT Act. Touching a UK system from India engages the Computer Misuse Act. When you test systems hosted in other countries β€” cloud makes this the default β€” your authorisation must be valid under the law where the systems sit, not merely your own.

Authorisation in practice: the paperwork of an engagement

Professional testing runs on five documents, from heaviest to lightest.

The engagement letter is the master contract: parties, dates, legal basis, liability, insurance, confidentiality, and the signature of someone with authority to bind the client. Attached to it come the scope β€” the exact list of targets, addresses, applications, permitted techniques and depth limits. Anything unnamed is excluded; Chapter 1's technical block showed this. The rules of engagement are the operating manual inside the scope: working hours, escalation and emergency-stop procedures, evidence handling, data-destruction duties, and the client contacts who know the test is live.

The emergency-contact sheet names a reachable human on each side, 24 hours a day, for every day of the test. It is boring until the moment it is the most important document you own. Finally, the authorisation letter β€” often called the "get-out-of-jail letter" β€” is the one-page signed excerpt the testers carry physically: who they are, what they are authorised to do, where, until when, plus a phone number a police officer, guard or helpdesk analyst can call at 2 a.m. to verify it. The Coalfire case is what that letter is for. The gap in their engagement β€” the sheriff not being informed β€” is what the letter alone cannot fix. That is why advance notice to local authorities is a standing rule of physical-scope testing, not a courtesy.

When scope is unclear. You find a system one hop from an in-scope address. The scope does not name it. The plan did not expect it. It looks valuable. The professional protocol is three moves, always the same. Stop β€” do not touch the system. Ask β€” in writing, to the authorised client contact, quoting the unclear point. Document β€” record the question and the answer in the engagement log. An expansion agreed by email and countersigned is then appended to the scope. What you never do is resolve doubt in your own favour. "They probably meant to include it" is the sentence that ends careers.

πŸ”§ Technical

In plain words: rules of engagement are the operating manual inside the scope β€” who works when, what they may and may not do, whom to call, and how evidence must be handled.

A condensed rules-of-engagement document. Everything here is fictional β€” "NiveshCo" is a synthetic company, and every name and address is invented:

RULES OF ENGAGEMENT β€” RoE-2026-014 (NiveshCo ↔ Vetta Security)
1. Window: 12–23 Jan 2026, 09:00–19:00 IST, Mon–Sat.
2. Testers: P. Rao (ID VET-1042), S. Kaur (ID VET-1077). No substitutions
   without written amendment.
3. In-scope: portal.niveshco.lab.local; 192.0.2.10-20 (lab segment).
   Excluded: production DB, card-data stores, third-party SaaS, physical
   sites other than Floor 3 server room, social engineering of staff.
4. Prohibited techniques: DoS/stress testing; destructive exploits;
   persistence beyond proof; real customer data in evidence.
5. Escalation: Vetta lead β†’ NiveshCo security lead (R. Menon, +91-XX…)
   β†’ CISO. ANY safety incident: stop work, call, log timestamp.
6. Authorities notified in advance (physical/cloned-badge tests):
   building security ops; local police station diary entry ref. XX/2026.
7. Evidence: encrypted volume provided by client; screenshots redact
   personal data; all copies destroyed within 30 days of report
   acceptance, destruction certificate to client.
8. Ambiguity protocol: stop β†’ written question β†’ logged answer.
   Silence is NOT authorisation.

Three clauses worth memorising. Clause 6 exists because of the Coalfire gap. Clause 4 keeps proof-of-concept inside DPDP-respecting bounds. Clause 8 makes the stop-ask-document protocol contractual, not just a habit.

Bug bounties and coordinated disclosure

For most people, the first authorised target will not be a client's network. It will be a public bug bounty or vulnerability-disclosure programme. The mechanics are simple and strict. An organisation publishes a policy. The policy defines its scope (which properties may be tested), its rules (which techniques are allowed), its safe harbour (a written promise that research inside the policy is authorised and will not be met with legal action), and its disclosure expectations. A bounty programme adds payment for verified findings. A plain disclosure policy pays in thanks. Both grant the same thing that matters: authorisation, narrowly drawn.

The cardinal rule: the policy's edges are the authorisation's edges. Testing an acquired company the policy has not listed, probing a third-party service the target uses, or trying a technique the policy excludes β€” each takes you outside the safe harbour, back into the world of sections 43 and 66 or their foreign equivalents. When unsure, ask through the programme's channel first. Most programmes answer scope questions. A written "yes" costs a day. A wrong guess can cost everything.

Coordinated disclosure is the ethics of what happens after you find something. You report privately. The organisation fixes. Publication happens by agreement β€” commonly after a fixed window (90 days is a widespread convention), or earlier with consent, with enough detail to help defenders without assisting attackers who have not patched. Researchers do not set deadlines like threats. Organisations do not sit on fixes forever. Both sides put users' safety first. In India, vulnerabilities affecting Indian systems can also be reported to CERT-In's vulnerability-reporting channel, which coordinates with affected vendors. That is a civic act, and a professional habit.[7]

πŸ”§ Technical

In plain words: a disclosure timeline counts the days from your private report to agreed publication. It balances the user's need for a fix against the researcher's need for recognition.

How a healthy timeline usually runs:

Day 0    Private report filed with programme/tracker (full detail,
          evidence, suggested fix). Countdown starts.
Day 0–7  Triage: vendor acknowledges, assigns severity, may ask for
          clarification. Researcher responds promptly.
Day 7–60 Fix development; researcher tests the fix in pre-release if
          offered (retest cooperation).
Day 90   Default coordinated publication: advisory published by both,
          credits the researcher, omits weaponisable detail until
          patch adoption spreads.
Extensions: granted for good cause (complex fix, dependency chains),
recorded in the tracker. Publication is never used as a threat,
and never withheld indefinitely as a favour.

The 90-day figure is a convention, not a statute β€” Google's Project Zero made it popular. But the structure (private report β†’ fix window β†’ coordinated advisory) is the professional norm worldwide. The Part C case-study chapters (25–26) show what happens when disclosure is handled badly instead.

Professional ethics: the rules you keep when nobody is checking

Certification bodies bind you to codes of conduct before they put their letters after your name. ISC2's code β€” binding on CISSP holders β€” requires protecting society, acting honourably and legally, providing diligent service, and advancing the profession. EC-Council's code for CEH holders likewise demands lawful, contracted work and confidentiality.[18] Behind the formal texts sit three working disciplines.

Confidentiality. Everything you see in an engagement β€” findings, architecture, credentials, the existence of the engagement itself β€” is client-confidential. Do not discuss it in interviews beyond what the client permits. Do not reuse snippets "because they're generic". Decline follow-on work that would require using knowledge of one client against them.

Handling what you encounter. Tests surface personal data, sometimes by accident, sometimes in bulk. The ethical protocol is minimisation. Look only as far as needed to prove the finding. Record only metadata: counts, patterns, and at most one redacted example. Never copy bulk data. Report the exposure so it gets closed. Copying a customer table "as evidence" is a DPDP problem and an IT Act sections 72/72A problem. It is also the fastest way to turn a successful test into a career-ending one.

Conflicts of interest. Declare them before work begins: shares in a competitor, family employed at the client, earlier jobs whose knowledge you would be selling back. The rule is disclosure, plus stepping aside where needed β€” not silence and hope. An auditor who cannot be independent is not an auditor. The same holds for a tester.

πŸ”§ Technical

In plain words: mandatory log retention creates a rolling window into the past. Investigators can only prove what the logs still remember. That is why defenders keep logs, and why the retention number matters in court.

Three effects of the 180-day CERT-In rule worth fixing in mind:

  • The window effect. An intrusion found on day 200, with only 180 days of logs, may have its earliest foothold already erased. The investigation rebuilds from the oldest surviving record and must state the horizon as uncertain.
  • The corroboration effect. Retention must survive rotation and reconfiguration. A migrated server that starts logging afresh silently resets the window. DFIR procedures check log continuity, not just log presence.
  • The locality effect. The Directions require retention within India. This shapes architecture β€” where Security Information and Event Management (SIEM) storage lives β€” and cross-border engagements alike. Cloud customers meet this fact during audits.

Chapter 32 turns this into practice: what a SOC keeps, where, and for how long.

Worked demo: anatomy of a paperwork failure

Return to the opening case and do what professional firms actually did after it: read the public record, then ask what each document should have said. The facts below come from the cited reporting.[1][2] The analysis that follows is ours, marked clearly where it goes beyond what reporting established.

What the record shows. The Judicial Branch of Iowa commissioned Coalfire to assess the electronic and physical security of courthouses, including Dallas County's. The testers worked under a contract and carried authorisation. The county sheriff's office had not been informed. On the evening of 11 September 2019 a deputy found the testers in the courthouse after hours. They were arrested on felony burglary and tools charges, held on US$100,000 bail each for about twenty hours, then released. The state admitted miscommunication. Felonies were reduced in February 2020. Everything was dropped by August 2020. In January 2026 the county settled the wrongful-arrest suit for US$600,000.

Clause by clause β€” what the paperwork should have said. First, the scope should have listed the physical sites building by building: "Dallas County Courthouse, Adel, after-hours entry permitted". Then no party could later describe the testers' presence as unexplained. Second, the rules of engagement should have had a notification clause obliging the client to brief every authority with jurisdiction over each site before its window opened β€” sheriff's office, building security, any alarm-monitoring company β€” and to log that briefing with a reference number. Third, the emergency-contact sheet should have named a client representative reachable at 9 p.m., with the standing to tell a deputy "yes, this is ours" before an arrest decision was made. Fourth, the carried letter should have been written for its real audience: a police officer at night, not a lawyer at noon. Plain language, the site names, the client's seal, and that phone number.

Where analysis parts company from record. Reporting does not tell us which of these clauses existed in the real contract, or exactly what the sheriff's office had been told by whom. What the record does establish is the outcome: valid authorisation failed to protect two professionals, because the people on the ground stood outside its chain of communication. That is the finding to design against. It is why the template in this chapter's core lab builds notification and 2 a.m.-readability into the document from the first draft.

Authorisation
Informed, competent, specific written permission to test defined targets. The fact that separates testing from crime.
Engagement letter
The master contract for a test: parties, legal basis, liability, confidentiality, signatures.
Scope
The listed targets, techniques, depth and time window of a test. Anything unnamed is excluded.
Rules of engagement (RoE)
Operating rules inside the scope: hours, escalation, emergency stops, evidence handling, notification duties.
Authorisation letter ("get-out-of-jail letter")
The one-page signed carried proof of authorisation, written for a doubtful officer at 2 a.m.
Ambiguity protocol
Stop, ask in writing, document the answer. Never resolve scope doubt in your own favour.
IT Act sections 43 / 66
India's civil-criminal pair for unauthorised access: compensation under 43; up to three years or β‚Ή5 lakh or both under 66, when done dishonestly or fraudulently.
Section 66F
Cyber terrorism under the IT Act. Punishable with imprisonment for life.
DPDP Act, 2023
India's digital personal-data-protection law: notice, consent, security safeguards, breach intimation. Penalties up to β‚Ή250 crore. Rules notified November 2025, phasing in.
CERT-In Directions, 2022
Binding Indian directions: report listed incidents within six hours; keep logs 180 days within India.
CFAA
The US Computer Fraud and Abuse Act (18 U.S.C. Β§1030), narrowed by Van Buren (2021) to a gates-up-or-down reading of authorisation.
Safe harbour
A programme's written promise that research inside its policy is authorised and will not be pursued legally.
Coordinated disclosure
Private report β†’ fix window (90-day convention) β†’ joint advisory, with users' safety governing both sides.
Conflict of interest
Any stake that could bias your testing or reporting. Managed by disclosure and, where needed, stepping aside.

Summary

Law and ethics are not the fine print of security work. They are its operating system. Written, informed, competent, specific authorisation is the line between a career and a criminal record. The Coalfire courthouse case shows that even genuine authorisation fails when it does not reach the people enforcing the law on the ground. India enforces the line through the IT Act's civil-criminal pair (sections 43 and 66), its named offences (66B–66F, 72, 72A), the DPDP Act's consent-and-penalty regime now activated by the 2025 Rules, and CERT-In's six-hour reporting and 180-day log rules. Banking and markets add RBI's and SEBI's frameworks. Globally, the CFAA (narrowed by Van Buren), the UK Computer Misuse Act and the EU's GDPR and NIS2 draw parallel lines with cross-border reach. The professional's armour is a small stack of documents β€” engagement letter, scope, rules of engagement, emergency contacts, carried authorisation β€” plus the stop-ask-document protocol for unclear scope, disclosure discipline in bounty work, and the ethics of confidentiality, minimal data handling and declared conflicts. Learn the paperwork as seriously as the techniques. Every later chapter of this book assumes you have.

Lab β€” Core tier: write the paperwork twice

⚠ Authorized practice only β€” this exercise, like every exercise in this book, runs only on systems you own or are explicitly authorised to test. Read Legal & ethical use first.

Objective: produce two written pieces β€” a post-mortem fix and a fresh authorisation. Use only this chapter and a word processor. Both are entirely fictional. Use synthetic names and placeholder addresses only.

Part A β€” what should the paperwork have said?

  1. From the worked demo's four-clause analysis, write the two clauses you judge most important for the courthouse scenario: the site-level scope line and the authority-notification clause. One paragraph each. Write them so a non-technical sheriff's captain could read them without explanation.
  2. Write the 2 a.m. paragraph of the carried authorisation letter: who the testers are, what they may do, where, until when, and the number to call. Six sentences at most.

Part B β€” draft a one-page authorisation and scope

  1. Invent a synthetic client. For example: a fictional regional fintech, "SahyogPay Systems", commissioning a web-application test of one lab-only property (portal.sahyogpay.lab.local) for ten working days.
  2. Fill the template below into a complete one-page document. Replace every bracket. Keep it to one page β€” that limit is the point.

Template

AUTHORISATION AND SCOPE β€” Ref [SAH-2026-XX]
[Client legal name] ("Client") authorises [Firm legal name] ("Firm") to
conduct a security assessment as follows.
1. Scope (exhaustive): [exact systems/URLs/IPs β€” placeholders only],
   techniques [permitted classes], depth limits [proof-of-concept rules,
   data-handling constraints].
2. Window: [start date] to [end date], [hours] [time zone], [days].
3. Named testers: [names, IDs]. No substitution without written amendment.
4. Exclusions: [named out-of-bounds systems/techniques β€” anything not
   listed in clause 1 is excluded by default].
5. Notification: [who at the Client and which external authorities are
   briefed before the window opens, with reference numbers].
6. Emergency contacts: Client [name, 24-h phone]; Firm [name, 24-h phone].
7. Evidence: [storage, redaction, destruction schedule].
Signed for the Client (authority to bind): ______________  Date: ______
Firm acknowledgement: ______________  Date: ______

Expected result

Part A: two clauses and a carried-letter paragraph a night-shift officer
could act on β€” plain language, site names, phone number, dates.
Part B: a completed one-page document in which every clause is specific
enough to enforce, exclusions are explicit, notification names real
roles (not "relevant stakeholders"), and the signature block names
someone with authority to bind the client. Then swap documents with a
peer: try to find one act of testing your partner's document permits
that its exclusions should have caught.

Lab β€” Technical tier: read a real policy, classify the edge cases

⚠ Authorized practice only β€” the hands-on tier runs strictly inside your own isolated lab VMs (Chapter 5) or a legal practice platform. Never a real third-party system. Legal & ethical use.

Switch to Technical mode for the hands-on lab.

Objective: reading and analysis only. You will read one real, public bug-bounty policy and use it to classify six fictional testing ideas. This is the exact judgement you will use the day you join a programme.

Steps

  1. Open Google's Vulnerability Reward Program rules and read them end to end: bughunters.google.com β€” Google and Alphabet VRP Rules.[17] Note specifically: how scope is defined, which behaviours are excluded, and how the policy treats services Google uses but does not own. Policies change. Classify against what the page says today.
  2. Classify each fictional idea below as in scope, out of scope, or ask first. Cite the policy section that decides it. All scenarios are hypothetical. Perform none of them.

Fictional scenarios to classify

  1. Finding reflected cross-site scripting in a listed Tier-1 Google web application's search parameter, tested with your own account.
  2. Brute-forcing your own Gmail account's password from your own script, to "test rate limiting".
  3. Finding a flaw in a small third-party support chat widget embedded in a Google page, and testing it on the live widget to confirm impact.
  4. Running a high-volume automated scanner against a listed application during business hours, "to cover more ground".
  5. Finding a weakness on the site of a company recently acquired but not yet listed in the programme's scope table.
  6. Emailing a Google employee a crafted link (phishing-style pretext) to show an account-takeover chain.

Expected result β€” answer key

1. In scope β€” listed asset, standard technique, your own account.
2. Out of scope/invalid β€” attacking your own account is account abuse,
   not a product vulnerability; policies exclude such tests. Report
   rate-limiting concerns via the programme.
3. Ask first β€” third-party components are usually outside the
   programme's authority. Check whether the policy carves them out
   before testing. Often the right move is reporting to the vendor.
4. Out of scope β€” high-volume scanning risks service degradation,
   which programmes prohibit. Volume is not diligence.
5. Out of scope / ask first β€” unlisted assets are by default not
   authorised; the scope table governs. Ask via the programme channel.
   Do not test while waiting.
6. Out of scope β€” social engineering of employees is excluded from
   essentially every programme, including Google's. Showing the chain
   without the human element (a report describing the potential) is
   the compliant path.

Record your six classifications, each with the policy section it leans on. Note any place where the policy left you unsure. Those notes are what a good "ask first" question is made of.

Quiz

Twenty questions: fifteen core and five marked Technical. Pass mark 60 per cent. Every answer carries an explanation.

Further reading

  • The IT Act, 2000 (as amended) on India Code β€” read sections 43, 66, 66B–66F, 70, 70B, 72, 72A and 45 in the current text[3].
  • The DPDP Act, 2023 and the notified DPDP Rules via MeitY and PIB β€” the primary sources for India's data-protection regime[5][6].
  • CERT-In's Directions page β€” the annexure listing reportable incident types is worth reading once in full[7].
  • Shreya Singhal v. Union of India (2015) and Van Buren v. United States (2021) β€” two judgments that reshaped their statutes. Read the summaries on the courts' own sites[4][13].
  • Google's VRP rules β€” a well-kept public policy, worth studying as a model of clear scope language[17].
  • ISC2's Code of Ethics β€” short, binding, and a good template for your own professional rules[18].

References

  1. CNBC, "Iowa paid Coalfire to pen test courthouse, then arrested the pen testers", 12 November 2019 β€” cnbc.com.
  2. Dark Reading, "County Pays $600K to Wrongfully Jailed Pen Testers", 2 February 2026 β€” darkreading.com.
  3. India Code, Information Technology Act, 2000 (as amended by the IT (Amendment) Act, 2008) β€” indiacode.nic.in.
  4. Supreme Court of India, Shreya Singhal v. Union of India, judgment of 24 March 2015 β€” main.sci.gov.in (search judgments by title).
  5. Ministry of Electronics and Information Technology, Digital Personal Data Protection Act, 2023 β€” meity.gov.in.
  6. Press Information Bureau, Government of India, notification of the Digital Personal Data Protection Rules, 2025, November 2025 β€” pib.gov.in.
  7. CERT-In, Information Security Practices, Procedure, Prevention, Response and Reporting of Cyber Incidents, Directions No. 20(3)/2022-CERT-In, 28 April 2022 β€” cert-in.org.in.
  8. Mondaq, "6 Hours To Report Cyber Incidents To CERT-In And More", 13 January 2023 β€” mondaq.com.
  9. Reserve Bank of India, "Cyber Security Framework for Banks", circular DBS.CO/TCSEC/006/31.01.023/2015-16, 2 June 2016 β€” rbi.org.in.
  10. Business Standard, "Sebi releases new cyber security framework for regulated entities", 20 August 2024 β€” business-standard.com.
  11. AZB & Partners, "Clarifications and FAQs regarding Cybersecurity and Cyber Resilience Framework", 29 July 2025 (deadline extension to 31 August 2025) β€” azbpartners.com.
  12. Legal Information Institute, 18 U.S.C. Β§ 1030 β€” Computer Fraud and Abuse Act β€” law.cornell.edu.
  13. Supreme Court of the United States, Van Buren v. United States, 593 U.S. 374 (2021), decided 3 June 2021 β€” supremecourt.gov.
  14. United Kingdom, Computer Misuse Act 1990 β€” legislation.gov.uk.
  15. European Union, Regulation (EU) 2016/679 (GDPR) β€” eur-lex.europa.eu.
  16. European Union, Directive (EU) 2022/2555 (NIS2) β€” eur-lex.europa.eu.
  17. Google, "Google and Alphabet Vulnerability Reward Program (VRP) Rules" β€” bughunters.google.com.
  18. ISC2, Code of Ethics β€” isc2.org.

Unverified: check source Three items to verify at class time: the current status of IT Act s.43A in the DPDP transition; the exact phased start dates of the DPDP Rules, 2025; and the final compliance status of SEBI CSCRF deadlines for the entity class you care about. Statutory texts and deadlines move; the chapter's structure does not.