adopt

name	adopt
description	Use when bootstrapping an existing project into PTSD pipeline

Instructions

1. Run ptsd adopt --dry-run first to see what will be discovered (BDD files, test files, Go packages). Review the output before proceeding.
2. Run ptsd adopt to create .ptsd/ structure and register discovered features.
3. Run ptsd status --agent to verify what was created.

Post-Adopt Checklist

After adopt completes, work through these steps for each discovered feature:

1. Assess status. Run ptsd feature list --agent. Features from BDD files start as in-progress. Features from test files start as deferred -- activate them with ptsd feature status <id> in-progress when ready.
2. Choose pipeline profile. Default is standard (PRD -> BDD -> Tests -> Impl). Use ptsd feature pipeline <id> lite for simple features that already have tests but no BDD. Use full for data-heavy features that need seed data.
3. Map tests to features. For each test file, run ptsd test map --feature <id> <test-file> (lite pipeline) or ptsd test map <bdd-file> <test-file> (standard/full). Without mapping, PTSD cannot track test results per feature.
4. Write PRD anchors. Every feature needs a section in .ptsd/docs/PRD.md with <!-- feature:<id> --> anchor. Document what the feature does, acceptance criteria, edge cases.
5. Verify. Run ptsd validate --agent -- fix any pipeline violations before committing.

When the project has no tests

If --dry-run finds no test files:

• Adopt creates empty .ptsd/ structure with no features
• Add features manually: ptsd feature add <id> "Title"
• Start from PRD stage -- the pipeline will guide you through BDD and tests

Working with deferred features

Features discovered from test files get status deferred and pipeline lite. This means:

• They are excluded from pipeline checks until activated
• Activate when ready: ptsd feature status <id> in-progress
• The lite pipeline skips BDD -- write tests directly from PRD

Common Mistakes

• Running ptsd adopt without --dry-run first -- always preview what will be discovered.
• Forgetting to map test files -- without mapping, ptsd status shows TESTS:0 even if tests exist.
• Setting all features to the same pipeline instead of assessing each individually.
• Not writing PRD anchors -- features without PRD sections will block at the first gate.
• Rewriting working code to fit PTSD patterns -- adopt tracks what exists, it does not require refactoring.