Skip to content

learning-paths.md

Latest commit

 

History

History
98 lines (70 loc) · 2.68 KB

learning-paths.md

File metadata and controls

98 lines (70 loc) · 2.68 KB

Learning Paths

This guide helps you choose the shortest route to useful outcomes based on your current level and job.

Pick a Path

You are here Start with Time budget Finish line
First-day user quickstart.md 30-45 minutes Core checks pass and repo is runnable
Working contributor workflow-recipes.md 45-90 minutes You can run the right gate sequence for your task
Power operator power-user-playbook.md 60-120 minutes You can tune depth, cost, and evidence posture
Claude adapter user claude-code-adapter.md 30-60 minutes You understand activation behavior and guardrails

Path 1: First-Day User

Goal: move from clone to confidence without accidental chaos.

  1. Complete quickstart.md.
  2. Run:
make check
./scripts/dp review --json
  1. If either fails, use:
    • ../troubleshooting/common-issues.md
    • ../troubleshooting/diagnostics-checklist.md

Success criteria:

  • make check exits 0.
  • review JSON reports "ok": true.

Path 2: Working Contributor

Goal: ship feature or fix work safely.

  1. Read workflow-recipes.md.
  2. Use the recipe that matches your current job.
  3. Confirm final gate sequence:
make check
./scripts/dp verify --json
./scripts/dp enforce pre-commit --policy dp-policy.json --json
./scripts/dp enforce pre-push --policy dp-policy.json --json

Success criteria:

  • Command outputs and exit codes indicate pass.
  • You can explain what each gate protected.

Path 3: Power Operator

Goal: optimize correctness, cost, and throughput under real constraints.

  1. Read power-user-playbook.md.
  2. Add evidence-first workflow to daily practice.
  3. For high-risk work, add targeted gate runs:
make rustdoc-check
make proptest-gate
make claude-adapter-gate

Success criteria:

  • You choose assurance depth intentionally, not by habit.
  • High-risk claims are backed by artifact paths, not memory.

Path 4: Claude Adapter User

Goal: understand when adapter behavior is fast, deep, or intentionally conservative.

  1. Read claude-code-adapter.md.
  2. Validate adapter scenarios:
make claude-adapter-gate
  1. If behavior surprises you, inspect:
    • ../internals/ooda-and-execution.md
    • ../internals/runtime-walkthrough.md

Success criteria:

  • You can predict activation and mode decisions for common prompt types.
  • You can locate the runtime modules that produced a decision.

Suggested Progression

  1. quickstart.md
  2. workflow-recipes.md
  3. claude-code-adapter.md (if relevant)
  4. power-user-playbook.md
  5. ../reference/command-reference.md

If you skip steps, the project will not punish you. It will just do it indirectly through failed gates.