security: remove .claude/ directory from git tracking#14
Closed
realityinspector wants to merge 217 commits intomainfrom
Closed
security: remove .claude/ directory from git tracking#14realityinspector wants to merge 217 commits intomainfrom
realityinspector wants to merge 217 commits intomainfrom
Conversation
- Fixed Entity creation in workflows.py to use correct schema fields (temporal_span_start/end instead of timepoint) - Fixed malformed print statement in test_parallel_execution.py - Removed temporary test files after verification
- Added retry_with_backoff function with exponential backoff (1s, 2s, 4s delays) - Integrated retry logic into all LLM client methods (populate_entity, validate_consistency, score_relevance) - Added proper error logging and graceful failure handling - Tested retry mechanism with success, eventual success, and complete failure scenarios - Updated CHANGE-ROUND.md to reflect completion and mark Phase 1 as complete
- Added LLM-powered knowledge generation during resolution elevation - Entities gain 3-12 new knowledge items when elevated to SCENE/DIALOG/TRAINED levels - Creates exposure events for newly learned knowledge - Added mock knowledge generation for testing in dry-run mode - Updated ResolutionEngine to accept LLM client for knowledge enrichment - Updated all ResolutionEngine instantiations to pass LLM client - Tested knowledge growth from 3 → 14 items across two elevation steps - Updated CHANGE-ROUND.md and README.md to reflect completion
- Added ModelManager class to fetch and cache Llama models from OpenRouter - Updated LLMClient to use Llama 70B (meta-llama/llama-3.1-70b-instruct) as default - Added model selection capability with configuration support - Added interactive model management CLI mode (python cli.py mode=models) - Added 'models' command to interactive query interface - Updated all hardcoded model references to use dynamic model selection - Added 24-hour caching for model lists to avoid excessive API calls - Enhanced CLI with model testing, switching, and detailed info features - Updated configuration to support model selection and cache TTL settings - Maintained backward compatibility with existing functionality
- Replaced OpenAI client with custom OpenRouterClient using httpx - Removed instructor dependency for structured outputs, handling JSON parsing manually - Updated QueryIntent to inherit from BaseModel instead of OpenAISchema - Added httpx>=0.27.0 to requirements.txt for HTTP client functionality - Removed openai>=1.57.0 from requirements.txt - Updated all LLM methods to handle JSON parsing without instructor - Enhanced prompts to explicitly request JSON output format - Maintained full functionality: entity population, validation, query parsing, relevance scoring - All tests pass, model selection works, knowledge enrichment functions properly - System now uses only httpx for OpenRouter API calls (no OpenAI SDK)
Answers 5 integration questions: connection security, health checks, data security, document accuracy, and suite rigging. Point-in-time audit covering all 13 Timepoint Suite repos.
- MECHANICS.md: "18 mechanisms" → 19, add verified cyclical template - QUICKSTART.md: cyclical mode now has verified template (agent4_elk_migration) - EXAMPLE_RUN.md: castaway_colony entity count 8 → 10 - dashboards/README.md: STANDARD → actual mode names, fix resolution level names
- Remove NONLINEAR from TemporalMode enum (orphaned, no templates use it) - Delete GroqProvider (dead code, never instantiated at runtime) - Clean run.sh: remove 28 orphaned template shortcuts (core/stress/portal), update help text from 41 to 21 templates - Delete TIMEPOINT_PRO_SUITE_INTEGRATION_ANALYSIS.md (temporary doc, served its purpose) - Document isolation-by-design: no Flash/Billing/Clockchain dependencies - Document planned /api/data-export endpoint for SNAG-Bench Axis 2 and Proteus - Note Pro-Cloud boundary: /api/usage and /api/budget not in public repo
ModelSelector.select_model() is the M18 implementation but was never tracked in mechanism_usage. Every other mechanism (M1-M17, M19) has tracking; M18 was the only blind spot.
…tion docs - Add training_data_unrestricted field to ModelProfile (M18) - Llama models marked restricted: outputs cannot train non-Llama models - Add for_training_data param to select_model() for MIT/Apache-2.0 filtering - Add get_training_safe_models() helper - Add examples/sample_training_data.jsonl (3 PORTAL mode examples) - Add "Cloud Execution via Pro-Cloud" section to README - Add "Sample Training Data" section to README - Update all docs with clear training data licensing requirements - Whitelist examples/*.jsonl in .gitignore
The "core" category doesn't exist; valid categories are showcase, convergence, persona, portal. Changed to --category showcase.
Remove VR references. Add Rendered Futures section, M20 planned mechanism, TDF interoperability, agent swarm use cases, PoCC protocol reference, and unified suite context table with whitepaper reference.
Three data-rich visual elements following Tufte/Knaflic/Bertin principles: - RAG vs SNAG comparison table: instant conceptual anchoring for AI practitioners - SNAG pipeline flowchart (mermaid TD): makes the engine pipeline concrete - Suite flywheel diagram (mermaid LR): visualizes the circular reinforcement loop All render on GitHub mobile. No chartjunk. Each element has one job.
Replace hand-rolled TDFExporter with the shared timepoint-tdf library
(from_pro + write_tdf_jsonl) so wire format stays consistent across the
Timepoint suite. Fix /api/data-export/{run_id} to return full payload
(entities, dialogs, causal_edges, metadata) instead of metadata-only.
feat: integrate timepoint-tdf package for TDF serialisation
- Enforce branch protection on main (1 approval, dismiss stale, no force push) - Ruff auto-fix: 3898 lint issues resolved, 246 files reformatted - README: add 2026-03-02 changelog section (TDF integration + protection) - Dev cruft removed (.pyc, __pycache__, .DS_Store, .pytest_cache) - Secret scan: no leaked credentials found
…3-02 chore: branch protection locked + ruff lint pass + security scan
Adds docs.json with Timepoint logo config to override default Mintlify discovery branding (was showing X AI logo).
fix: add Timepoint branding for Mintlify docs
chore: rename org timepoint-ai → timepointai
Pulls in TDF v1.2.0 which adds clockchain schema versioning and model provenance fields (text_model, image_model, model_provider, model_permissiveness, schema_version, generation_id) to TDFProvenance. The from_pro() and from_clockchain() transforms handle the new fields automatically — no Pro-side code changes required.
chore: update timepoint-tdf to v1.2.0
Disable default /docs, /redoc, /openapi.json endpoints and re-mount them behind get_current_user auth dependency so API schema is not publicly exposed.
feat(api): gate OpenAPI docs behind API key auth
Remove private repo table entries (Web App, iPhone App, Billing, Landing), Cloud Execution via Pro-Cloud section revealing Railway/internal service details, and Pro-Cloud boundary references with internal env var names. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- README: add Cloud Execution section with subdomain refs, update Suite table with subdomain URLs (flash/clockchain/pro/proteus.timepointai.com), remove private repo names and Railway internal references - AGENTS: add subdomain architecture table, OpenAPI docs section, public repo security rule - FRONT-END-SPEC: new file documenting dashboard API endpoints, auth flow (Flash SSO via token introspection), query parameters, CORS policy, and planned UI surfaces
docs: subdomain architecture, OpenAPI docs, security cleanup
Adds .claude/ to .gitignore and removes all tracked .claude/ files from version control. Fixes #13
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Removes the
.claude/directory from git tracking and adds it to.gitignore.The
.claude/directory contains local Claude Code configuration files (commands, agents, settings) that should not be committed to version control.Changes:
.claude/to.gitignore(replacing the previous partial entries for.claude/settings.local.jsonand.claude/*.log).claude/files from git tracking viagit rm --cachedFixes #13