Run Project 1 release candidate suite

1. 01

   Concepts & system design

   60 min· read · map the mechanismtimer

   R28 — PyRIT evaluation

   R31 — NIST AI RMF documentation

   R42 — GenAI release evidence

   Read the assigned sections; map the mechanism, threat model and trade-offs.

2. 02

   Project implementation

   120 min· build · integrate · debugtimer

   Execute the complete P1 Release Candidate suite: unit/integration tests, D1 Crescendo multi-turn fuzzer, D4 PII gateway checks, D11 trajectory ledger verification, D12 SBOM scan, and D8 PyRIT CI gate.

   Build, integrate, debug and improve the increment in the active project.

3. 03

   Security lab

   60 min· attack your own systemtimer

   No separate task text for this block on day 79 — the sourced work sits in the block above. Keep the time box.

   Test the abuse case, failure path, authorization bypass or resource limit.

4. 04

   Verification & evidence

   30 min· test · record evidencetimer

   Freeze P1 RC package: architecture diagram, STRIDE threat model, test report, model manifest, operational limits, and residual-risk statement.

   Add or run regression tests; record the 9-point evidence record.

5. 05

   Deliberate practice & Socratic review

   30 min· explain aloud · update gapstimer

   Answer the 8 Socratic questions without notes: What? Why? How? What if? Why not? How can I prove it? Where can it break? What have I learned?

   Re-explain the hardest concept aloud, inspect one edge case, update the gap log.

6. total5 h

• R28Microsoft PyRIT↗ github.com

  Python Risk Identification Tool: targets, orchestrators, converters, scorers, and multi-turn attack automation.

• R31NIST AI RMF 1.0↗ nist.gov

  Govern, Map, Measure, and Manage functions; risk ownership, measurement, and lifecycle documentation.

• R42NIST AI RMF GenAI Profile↗ doi.org

  NIST AI 600-1: generative AI risks and suggested actions for confabulation, privacy, information integrity, and value chain.

Read only the named sections needed for today's task. Reading sits inside the 60-minute concepts block; it is not extra time.

• docsNIST (NIST AI 600-1) · framework profile
  Artificial Intelligence Risk Management Framework: Generative Artificial Intelligence Profile

  Maps suggested actions to GV/MP/MS/MG subcategories for confabulation, information integrity, and value chain risk, giving the release-candidate report its evidence taxonomy.

  R42R31checked 2026-11-11
• repoMicrosoft Azure (GitHub) · framework + docs
  PyRIT

  Provides targets, orchestrators, converters, and scorers, so the D1 Crescendo fuzzer and the D8 CI gate are configured runs of the same primitives rather than bespoke scripts.

  R28checked 2026-11-11
• videoMicrosoft Developer · 10:59
  Episode 10: Automating Multi-Turn Attacks with PyRIT | AI Red Teaming 101

  Runs PyRIT's Crescendo and Adversarial Chat orchestrators against a live target with objectives and scoring, which is precisely the multi-turn fuzzer that must pass before release.

  R28checked 2026-11-11
• videoMicrosoft Developer · 8:12
  Episode 8: Automating AI Red Teaming with PyRIT | AI Red Teaming 101

  Explains PyRIT's architecture — targets, scoring engines, and memory — before the demo, which is what you need to wire a deterministic CI pass/fail threshold instead of manual review.

  R28checked 2026-11-11
• articleUSENIX Security 2025 · paper + open access
  Great, Now Write an Article About That: The Crescendo Multi-Turn LLM Jailbreak Attack

  Defines the escalation mechanism (benign opening turns that reference the model's own prior output) and reports per-model attack success rates, giving the D1 fuzzer its objective and baseline.

  R28R42checked 2026-11-11

A lab is complete only when it produces evidence. Installing PyRIT, Garak, Presidio or TransformerLens is not completion. Saved on this device only.