Consolidated Security Portfolio

1. 01

   Concepts & system design

   60 min· read · map the mechanismtimer

   R28 — PyRIT

   R31 — NIST AI RMF

   R34 — TransformerLens limits

   R35 — NVD

   R42 — NIST GenAI Profile

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

2. 02

   Project implementation

   120 min· build · integrate · debugtimer

   Assemble the top-level Master Portfolio Index linking: the 3 flagship repos, the 12 embedded security deliverables (D1–D12), the implemented agentic capabilities, the S12 public CVE teardown report, the S10 mechanistic interpretability reproducibility report, the D8 CI/CD security gate logs, and the D12 supply-chain SBOM manifests.

   Conduct a closed-book 90-minute oral defense aloud: explain the architecture, design alternatives, failure modes, attack reproductions, and residual risks for P1, P2, and P3 (D1–D12). Record any hesitation in your gap log and create a prioritized post-555h continuation plan (e.g., GPU/vLLM inference benchmarking, TTFT/ITL/KV-cache optimization, and targeted outreach).

   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 111 — 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

   Published Master Portfolio Repository + recorded oral defense notes. 555-Hour Roadmap Complete!

   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.

• R34TransformerLens Docs↗ transformerlensorg.github.io

  HookedTransformer, residual stream, ActivationCache, Logit Lens, and Activation Patching tutorials.

• R35NIST NVD & CISA KEV↗ nvd.nist.gov

  CVE records, CPE affected-version ranges, CVSS v3/v4 scoring context, and Known Exploited Vulnerabilities catalog.

• 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.

• docsTransformerLensOrg (Neel Nanda) · Docs + tutorial index
  TransformerLens Documentation

  Covers HookedTransformer, the residual stream, ActivationCache, Logit Lens, and activation patching, the toolkit your S10 reproducibility report uses to make a causal claim about a specific head or layer.

  R34checked 2026-11-11
• videoInternational Association for Safe & Ethical AI · 25 min
  An Introduction to Mechanistic Interpretability – Neel Nanda | IASEAI 2025

  Frames what interpretability can and cannot establish today, including the limits of current methods, which is exactly the honesty bar for a reproducibility report you will defend aloud.

  R34checked 2026-11-11
• videoFIRST · 58 min
  CISA’s Known Exploited Vulnerabilities (KEV) Catalog

  Two CISA analysts explain what KEV is and is not, how a CVE gets added, and how it feeds SSVC prioritisation — the exploitation signal that outranks raw CVSS in your portfolio's risk story.

  R35checked 2026-11-11
• articleFOSSA · 10 min read
  Using the CISA KEV Catalog

  Lays out the three KEV inclusion criteria and a layered triage order (direct dependency, CVSS floor, fix available, KEV, then EPSS), which is a defensible method to cite for D12 findings.

  R35checked 2026-11-11
• labTransformerLens Documentation · Runnable notebook
  Transformer Lens Main Demo Notebook

  Runs run_with_cache, then activation patching on the IOI task with a clean/corrupted metric, producing the layer-by-position patching grid you need to reproduce and defend a circuit claim.

  R34checked 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.