# Breach Lab: instructor notes Breach Lab is a quiz for students who start cyber security studies. Game studies students can use it to examine feedback, scoring, accessibility, and learning design. The arcade interface has been removed. Previous releases remain in Git history. ## Start and practise Open `/learn.html` and select Play. The city journey starts immediately. The first practice question opens at its rooftop terminal. Each submitted answer has an explanation. Select Next question to continue directly. Practice has no timer, no failure limit, hints, and retries. Use the field kit to decode text and compute SHA-256. Firewall questions include a traffic test. A worked solution becomes available after three errors and marks the task as assisted. A practice run has eight questions with increasing difficulty. The library has 40 tasks and lets students select a single topic and tier. A route code can reproduce a practice sequence. Reports and learner notes stay in the tab unless the student downloads a report. Reloading the page clears the local run. ## Monthly Top 10 The monthly challenge is optional. It has eight questions and one accepted answer per question. All learners receive the same ordered question set for that calendar month. The set is frozen in the database when the month first starts. Later edits to the practice bank do not change that month's challenge. A new month uses a new set. Month boundaries use Europe/Helsinki time. Correct-answer count is the primary ranking. Shorter total answer time breaks ties. The timer starts on the server when it sends a question and stops when it accepts the answer. Network delay is included. It does not count explanation reading time. The next question starts timing only when the learner selects Next question. The timer continues when the learner leaves the tab. The session expires after two hours or at the next month boundary. Completing the challenge does not automatically post a score. A learner can select Post to monthly Top 10 on the result page. The public row contains a generated call sign, correct-answer count, and answer time. It contains no student name. One run can post only one row. Repeated runs are allowed and get different call signs. The board ranks runs, not verified individual people. The server validates answers and ignores client-supplied scores and times. Sessions use random bearer tokens held in tab memory. The API does not send correct answers with an unanswered question. The practice bank still contains its answer data, and learners can repeat or automate runs. This board is a learning activity, not evidence of independent assessment. Do not use it as a course grade. ## Concepts and model limits | Topic | Student action | Key distinction | | --- | --- | --- | | Codes and encoding | Decode hex, Base64, binary, Caesar, and nested formats | Encoding does not provide secrecy | | Network boundaries | Reorder rules and test traffic | The first matching rule wins | | Identity and access | Select factors, permissions, and account responses | Authentication differs from authorization | | Hashes and integrity | Compute and compare digests | Integrity depends on a trusted reference | | Logs and evidence | Correlate events and normalize time | An observation differs from an inference | | Secure code | Select safe APIs and server checks | Keep data separate from instructions | | Incident response | Select containment, evidence, and recovery actions | Preserve evidence and record uncertainty | The fictional firewall evaluates inbound TCP packets by exact source address and port. `*` matches any source or port. Unmatched packets are dropped. It does not model connection state, IPv6, CIDR ranges, NAT, routing, application inspection, or a specific firewall product. The SHA-256 tool hashes exact UTF-8 text without adding a newline. It is not a password storage or signing system. No task sends traffic to a target system or executes submitted code. ## Data and operation Practice does not send answers or reports to the API. Monthly challenges send answers for checking and timing. Session records expire after two hours. Posted scores are kept for up to 90 days. The current month's Top 10 is public. Older rows are not exposed by the scoreboard endpoint. Expired records are removed when the API is used. No application account, cookie, localStorage, or sessionStorage is used. Basic rate limits are kept in service memory. NGINX and the reverse proxy can have their own access logs. This document does not define the infrastructure log retention policy. Back up the service SQLite database before future service changes. It contains posted results and frozen monthly questions. No test scores should be posted to the production board. Browser tests permit automated posting only to the isolated local preview database. ## Reference material - [OWASP: password storage](https://cheatsheetseries.owasp.org/cheatsheets/Password_Storage_Cheat_Sheet.html) - [OWASP: SQL injection prevention](https://cheatsheetseries.owasp.org/cheatsheets/SQL_Injection_Prevention_Cheat_Sheet.html) - [OWASP: injection prevention](https://cheatsheetseries.owasp.org/cheatsheets/Injection_Prevention_Cheat_Sheet.html) - [MDN: textContent and innerHTML](https://developer.mozilla.org/en-US/docs/Web/API/Node/textContent) - [RFC 5737: documentation address blocks](https://www.rfc-editor.org/info/rfc5737/) ## Cyber-city platform journey Play starts the side-view journey. The player controls the human white hat character with A / D or Left / Right. Learners must jump two roof gaps to reach each quiz terminal. There are eight quiz stops. Movement and city scrolling stop while the question and explanation are open. Next question opens the next section. Movement starts only when the player uses a direction control. Jump controls: Space, Up, W, or the up button. On touch screens, hold a direction and tap the up button with another finger. Escape or Pause stops the scene. Leaving the browser or hiding the tab also pauses it. Each run has five lives across all eight sections. A fall costs one life and returns to the last roof checkpoint. Movement stops after a fall. At zero lives, the player can restart or view the quiz report. Restart clears quiz progress and restores five lives. Later sections have wider gaps and smaller landing areas. There are no jump timing prompts. There are no monsters, patrols, weapons, or combat scores. The platforming tests the jump timing needed to reach a question; it is not a graded cyber security skill. The monthly challenge remains separate and quiz-only. Its timing rules are unchanged. The task library is available for students who want a direct exercise without platforming.