Previously the session-end protocol told the AI to update profile.md
if new insights emerged, with no guidance on where. Over many sessions
this caused freeform bold-text entries to accumulate without structure.
Now requires each update to match an existing H2 section in the file,
with explicit rules against creating new top-level sections.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Moves all customization commands (persona, modality, structure, update,
import) to a separate file that gets copied to .therapy/ during setup.
This allows existing users to receive new commands via the update flow.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Imports now create session files with original dates instead of just
reading once and extracting to profile. This provides natural relevance
decay as new sessions accumulate. Also adds post-setup import command.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Consolidates inner-dialogue-deeper content into the main repo. Removes
expansion pack gating, Gumroad references, and the two-tier setup flow.
All communication styles and therapeutic modalities are now included and
offered directly during setup. Bumps kit_version to 2.0.0.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Core personas should be Warm 4o-Style and Direct & Challenging.
Warm & Supportive is an expansion pack persona.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Update all references from "AI Therapy Starter Kit" to "Inner Dialogue",
rename expansion pack to "Inner Dialogue: Deeper", and update GitHub URLs
to ataglianetti/inner-dialogue.
CLAUDE.md (setup) was already reworked for core/expansion split.
This completes the remaining two files:
CLAUDE.template.md (in-session behavior):
- Modality switching checks installed modalities instead of hard-coding all 6
- Natural language triggers point to library dirs instead of listing expansion content
- Persona selection reads library and conditionally shows expansion options
docs/GETTING-STARTED.md:
- Setup overview shows core defaults (2 personas, CBT only)
- Customization sections reference Expansion Pack for additional content
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Captures the emotional warmth users loved from 4o while providing better
therapeutic technique. Feels like a good friend who asks insightful questions
rather than a therapist. Uses casual language, emoji, and disguised techniques.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Users can now delete the ai-therapy-kit repo after setup. Key changes:
- Setup copies ALL components to .therapy/library/ for local customization
- Customization reads from library/ instead of source_repo
- Updates fetch directly from GitHub via WebFetch
- Added natural language triggers for customization requests
- Added discoverability hints (first-session closing, help response)
- Removed source_repo from version.json (only source_url remains)
- Deleted obsolete setup.sh and setup.ps1 scripts
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Add docs/GETTING-STARTED.md for non-technical users
- Add "Known Limitations" section to README about LLM safety guardrails
- Add FAQ entry about AI seeming evasive on trauma topics
- Fix setup.sh prompt_choice outputting to stdout instead of stderr
- Fix first session detection in CLAUDE.template.md (check for empty
sessions folder before trying to reference previous sessions)
- Add guardrails acknowledgment instructions to CLAUDE.template.md
(AI should be honest when hitting built-in limitations)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Anthony Taglianetti