Skip to content
Roadmap0/111
not startedstudy day 3 of 6 · covers original day 3

Study Python types

P1 AssistantS0M1
not started0/5 blocks
0/5
Jump & context

Week 1: Application Foundations, Model Trust, Tools & Authorization Boundaries

The five hours

  1. 01

    Concepts & system design

    60 min

    R01 — Python config/dependency hygiene

    R10 — OWASP Secrets Management Cheat Sheet

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

  2. 02

    Project implementation

    120 min

    Study Python typing, exception handling, environment variables, secrets, and lockfiles. Add config loading and a clean LLM provider abstraction (LocalOllamaProvider vs. HostedProvider).

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

  3. 03

    Security lab

    60 min

    Inspect model source/revision and package dependencies; document why Hugging Face/Ollama model cards are metadata, not cryptographic proof of safety.

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

  4. 04

    Verification & evidence

    30 min

    Model & dependency inventory with pinned sources and versions; secrets verified outside Git.

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

  5. 05

    Deliberate practice & Socratic review

    30 min

    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

Read today · inside the concepts block

R01 R10

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

Free resources for this day

5 curated
  • docsOWASP Cheat Sheet Series
    Secrets Management

    Covers storage, rotation, and leakage paths for credentials, which is the rule set behind keeping hosted-provider keys out of your config module and repo.

    R10checked 2026-11-11
  • videoTech With Tim · 24:46
    Python Typing - Type Hints & Annotations

    Covers annotations, Optional/Union, and generics so the provider abstraction has a typed interface instead of a dict-shaped hole.

    R01checked 2026-11-11
  • videoBugBytes · 18:49
    pydantic-settings - Modern, Type-Safe Configuration for Python Apps

    Demonstrates BaseSettings, env prefixes, .env files, and validation on load, which is how config for LocalOllamaProvider vs HostedProvider gets typed and fails fast when a secret is missing.

    R01R10checked 2026-11-11
  • article12factor.net
    The Twelve-Factor App

    Factor III states config belongs in the environment, with the litmus test that the codebase could be open-sourced at any moment, which is the bar for separating secrets from code.

    R10checked 2026-11-11
  • repoGitHub
    pydantic/pydantic-settings

    Reference implementation of env-file and env-var settings sources, including nested config and secret types you can lift into the provider factory.

    R01checked 2026-11-11

Evidence record · 9-point standard

0/9

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

Day notes

empty

The two source formulations

verbatim · nothing summarised

Project & Stage Target

P1 | S0, M1

Core Implementation

Study Python typing, exception handling, environment variables, secrets, and lockfiles. Add config loading and a clean LLM provider abstraction (LocalOllamaProvider vs. HostedProvider).

Security Lab (Supply Chain)

Inspect model source/revision and package dependencies; document why Hugging Face/Ollama model cards are metadata, not cryptographic proof of safety.

Verification & Evidence

Model & dependency inventory with pinned sources and versions; secrets verified outside Git.

Read Today (60m Concepts Block)

R01 (Python config/dependency hygiene), R10 (OWASP Secrets Management Cheat Sheet).

---