Thanks to visit codestin.com
Credit goes to github.com

Skip to content

Conversation

@chrisquorum
Copy link

Summary

This PR adds Quorum Earth-specific customizations to the Agent Zero fork.

Changes

🤖 Specialized Agents (Commit 1)

  • Data_Architect: Dataset analysis and import planning
  • ProForma_Agent: Financial modeling and CFO-level analysis
  • Narrative_Agent: Pitch deck narrative writing
  • Khosla_Advisor: Pitch deck critique using Khosla methodology

🔧 MCP Integrations (Commit 2)

  • Sequential Thinking prompt for structured reasoning
  • Context7 prompt for documentation lookup
  • PRD protocol for product requirements
  • System prompt extensions for MCP tools

📁 Templates & Configuration (Commit 3)

  • GitHub automation templates (3-tier: minimal, standard, full)
  • Default and Quorum prompt profiles
  • API instrumentation helpers
  • Caching configuration

Stats

  • 39 files changed
  • 3,758 insertions
  • 3 commits

Testing

  • All agents tested in development environment
  • MCP integrations verified with Sequential Thinking and Context7 servers

chrisquorum and others added 30 commits January 24, 2026 16:44
Add four specialized agents for Quorum Earth pitch deck project:

- Data_Architect: dataset analysis and import planning

- ProForma_Agent: financial modeling and CFO-level analysis

- Narrative_Agent: pitch deck narrative writing

- Khosla_Advisor: pitch deck critique using Khosla methodology
Add Sequential Thinking and Context7 MCP integrations:

- Sequential thinking prompt for structured reasoning

- Context7 prompt for documentation lookup

- PRD protocol for product requirements

- System prompt extensions for MCP tools
Add GitHub automation templates and helper modules:

- GitHub templates (3-tier: minimal, standard, full)

- Default and Quorum prompt profiles

- API instrumentation helpers

- Caching configuration
- sync-upstream.yml: Weekly automated sync with agent0ai/agent-zero
- update-quorum.yml: Cascade sync to quorum branch
- CODEOWNERS: @ckantrowitz as default owner
…tification

- Add --repo flag to gh workflow run to target fork not upstream
- Remove 'Notify on failure' step (issues disabled on fork)
- Add Summary step for workflow visibility
- Add actions:write permission for workflow dispatch
- Skip update-quorum trigger in dry_run mode
fix: sync-upstream workflow repo reference and notification
- Create sync branch instead of pushing directly to protected quorum
- Check for changes before attempting merge
- Remove issue notification (issues disabled)
- Add proper summary output
- Handle merge conflicts gracefully with instructions
fix: update-quorum workflow to use PR-based approach
- Deploy FULL tier CI workflow with flake8, pylint, markdownlint, bandit, TruffleHog
- Configure workflow triggers for quorum and main branches
- Add .flake8, .pylintrc, .markdownlint.json linting configs
- Add .pre-commit-config.yaml for local development hooks
- Note: Existing codebase has linting issues to be addressed incrementally
- Add security patterns (.env, *.pem, *.key, secrets/)
- Add agent data exclusions (memory/, logs/, tmp/)
- Add Python, OS, IDE, and backup patterns
- Remove tracked .DS_Store files
chore: enhance .gitignore with comprehensive exclusion rules
Added missing CI environment skip to prevent test failures in GitHub Actions.
The file was missing the standard CI skip pattern that other test files have.

Pattern: if os.environ.get(CI) or os.environ.get(GITHUB_ACTIONS): pytest.skip()
Quorum Earth and others added 11 commits January 29, 2026 15:19
Restores send_file, jsonify, session, and request imports from Flask
that were removed by auto-linting. These imports are re-exported by
the api module and required by backup_create.py and image_get.py.

Fixes ImportError that prevented container startup.

Co-Authored-By: Claude Sonnet 4.5 <[email protected]>
faiss-cpu 1.11.0 depends on numpy.distutils which was removed
in numpy 2.0 and causes ImportError on startup. Upgrading to
faiss-cpu 1.13.2 resolves the compatibility issue.

Fixes ModuleNotFoundError: No module named 'numpy.distutils'

Co-Authored-By: Claude Sonnet 4.5 <[email protected]>
Updated SearXNG configuration based on official documentation:

limiter.toml:
- Replace deprecated [real_ip] with [botdetection]
- Add trusted_proxies list with localhost networks
- Fix configuration schema to match SearXNG requirements

settings.yml:
- Add missing engines to remove list (ahmia, torch, yacy images)
- Remove Ahmia blacklist plugin (engine doesn't exist)
- Clean up commented-out configuration options

Fixes all SearXNG startup warnings and deprecation notices.

References:
- https://docs.searxng.org/admin/settings/settings
- https://docs.searxng.org/admin/searx (limiter configuration)

Co-Authored-By: Claude Sonnet 4.5 <[email protected]>
Updates both default and hacker profile environment.md files with:
- Python virtual environment locations (/opt/venv-a0 vs /opt/venv)
- Venv activation guidance for Agent Zero dependencies
- Distinction between general-purpose and Agent Zero-specific venvs

Customization ID: C013
Add DDP routing matrix to behaviour_default.md for automatic subordinate delegation.

The protocol provides:
- Classification rules for request routing
- Routing matrix mapping intents to specialized profiles
- Precedence rules for multi-domain tasks
- Loop/depth safety guards
- Delegation message standards
- Verification obligations for delegators

Note: behaviour.md in memory/default/ is user data (gitignored).

Customization registry updated with C014 entry.
…mpts

Implements Run 3 patch to enhance delegation routing with:

- Added Task-to-Profile Routing Guide to call_subordinate tool prompt
  * Routing matrix table for 7 specialized profiles
  * Precedence rules for ambiguous cases
  * Clarification guidance (ask only when materially needed)
  * Delegation message template with 5 required elements

- Enhanced solving loop Step 3 with DDP classification
  * Delegation Decision Protocol integration
  * Quick-reference routing matrix
  * Explicit delegation rules and reset guidance

This creates three-layer routing architecture:
- Layer 1 (Behaviour): DDP classification and principles
- Layer 2 (Tool): Routing matrix and precedence rules
- Layer 3 (Solving): Execution-time routing guidance

Files modified:
- prompts/agent.system.tool.call_sub.md: 33 → 78 lines (+45)
- prompts/agent.system.main.solving.md: 27 → 43 lines (+16)
- CUSTOMIZATIONS.md: Added C015 entry

Aligns with Constitution (behaviour_default.md) from Run 2.
Add comprehensive delegation evaluation documentation suite:

- golden_routing_set.md: 13 canonical routing examples across 7 profiles

- adversarial_delegation_suite.md: 15 edge cases testing precedence rules

- validation_protocol.md: Repeatable 15-prompt test suite with success criteria

Updated docs/README.md with delegation evaluation section.

Added CUSTOMIZATIONS.md entry C016 for documentation tracking.
Add Run 4 validation testing documentation:

- validation_results.md: Complete test results with 15/15 tests passed (100% success)

- Documents routing accuracy, edge case handling, and protocol compliance

- Includes detailed analysis of each test case with actual vs expected routing

Updated CUSTOMIZATIONS.md with entry C017 for documentation tracking.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants