Skip to content

docs(community): add fork stewardship, contributing guide, and code of conduct#1088

Open
blackms wants to merge 40 commits intoruvnet:mainfrom
blackms:main
Open

docs(community): add fork stewardship, contributing guide, and code of conduct#1088
blackms wants to merge 40 commits intoruvnet:mainfrom
blackms:main

Conversation

@blackms
Copy link

@blackms blackms commented Feb 6, 2026

Summary

This PR adds baseline community governance documentation for the fork and clarifies project stewardship expectations.

What Changed

  • Add Fork Stewardship section in README.md
  • Add CONTRIBUTING.md with contribution workflow and upstream relationship guidance
  • Add CODE_OF_CONDUCT.md (Contributor Covenant 2.1)

Why

  • Improve contributor onboarding and expectations
  • Keep communication respectful and transparent
  • Document maintenance cadence rationale and collaboration principles

Scope

Documentation-only changes. No runtime or behavior changes.

blackms and others added 30 commits February 6, 2026 22:54
### What Changed
- add  section to README with clear, respectful fork rationale
- add  for fork contribution workflow and upstream relationship
- add  (Contributor Covenant 2.1)

### Why
- set transparent community expectations for maintenance cadence and collaboration
- provide standard open-source governance docs for contributors

### Notes
- documentation-only change; no runtime behavior changes

### References
- N/A
Replace hardcoded ruv co-author values in local settings and commit hook defaults for v2 and v3.
### What Changed
- add  to Anthropic provider model types and capabilities
- update v3 default model preferences and headless worker opus mapping
- add Opus 4.6 to integration router/provider catalogs and CLI provider listings
- update local Claude settings defaults to Opus 4.6
- refresh providers README supported model list

### Why
- enable runtime support for Anthropic Claude Opus 4.6 across v3 components

### Notes
- pricing and max output tokens aligned with official Opus 4.6 documentation used during implementation
- local build not executed because  is not installed in current environment

### References
- https://www.anthropic.com/news/claude-opus-4-6
- https://docs.claude.com/en/docs/about-claude/models/overview
- https://docs.claude.com/en/docs/about-claude/models/migrating-to-claude-4-6
### What Changed
- rewrote README and in-scope docs to keep only code-traceable claims
- added implementation-status/evidence/known-limitations sections across rewritten docs
- downgraded aspirational DDD and comparison claims where runtime evidence is partial
- added docs/AUDIT_DOCS_VS_CODE.md with adversarial verification report and gates

### Why
- enforce docs-as-code truth sync and eliminate doc/code drift

### Validation
- evidence pointers checked: 113/113 valid
- mermaid diagrams compiled: 2/2 via @mermaid-js/mermaid-cli
- no open contradictions and no critical findings in final audit
BREAKING FIX: hooksPostEdit, hooksPostTask, and hooksPostCommand were
returning success without actually storing any data. This fix makes
them call getRealStoreFunction() to persist to the memory database.

Changes:
- hooksPostEdit: Now stores to 'patterns' namespace with HNSW indexing
- hooksPostTask: Now stores to 'trajectories' namespace with HNSW indexing
- hooksPostCommand: Now stores to 'commands' namespace with HNSW indexing
- getLearningStats: Now queries real SELECT COUNT(*) instead of file_size/2KB

Before:
- handlers returned {recorded: true} without database INSERT
- hooksPostTask returned Math.random() for duration (fake data)
- statusline showed file_size/2KB as "patterns" count (misleading)

After:
- handlers call storeEntry() with generateEmbeddingFlag: true
- real pattern IDs returned from database
- statusline queries actual table row counts

Verified:
- ✅ post-edit: PERSISTED - "Pattern stored with HNSW indexing"
- ✅ post-task: PERSISTED - "Trajectory stored with HNSW indexing"
- ✅ post-command: PERSISTED - "Command pattern stored with HNSW indexing"

Closes ruvnet#1058

(cherry picked from commit 2e32f85)
docs: truth-sync README + docs + v3/docs with adversarial verification
fix(mcp): validate tool input before handler execution to prevent undefined-param crashes
…istence

fix(hooks): V3 MCP hook handlers now persist data (upstream ruvnet#1059)
…cleanup

fix(workers): cleanup background processes and avoid default background fan-out
fix(security): parameterize memory-initializer SQL queries
…doff-guard

fix(teammate): guard general-purpose subagent on legacy claude code
fix(hooks): fallback to hook payload when TOOL_INPUT vars are missing
…anup

chore(root): clean stale aliases and restore CI script map
chore(ci): fix rollback-manager CLI smoke path
…ract

chore(root): deprecate broken root entrypoint contract
chore(ci): fix stale publish command in v3-ci
…checks

test(docker): enforce strict MCP regression assertions
…checks

test(docker): remove permissive CLI regression fallbacks
- fix missing v2 package contract targets (main/start/files)

- restore neural/goal init script entrypoints

- replace undefined coverage script invocations

Refs: #47 #48 #49
* chore(v2): repair broken package/script contract surfaces

- fix missing v2 package contract targets (main/start/files)

- restore neural/goal init script entrypoints

- replace undefined coverage script invocations

Refs: #47 #48 #49

* docs(v2): remove stale npm script snippets

- replace dead script examples with runnable v2 scripts

- mark MCP registry steps as TODO placeholders instead of missing scripts

Refs: #50
- mark wrapper entrypoint files as deprecated internal paths

- add deprecation runtime notices with supported replacement command

- document staged removal plan and entrypoint contract

Refs: #51
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

2 participants