Hybrid multi-model task orchestrator — routes coding/review/scan tasks across LLMs via LangGraph
* chore: deprecation tombstone — README notice, remove stale security-review/ duplicate * chore: retrigger CI after removing go from CodeQL default-setup scope |
||
|---|---|---|
| .github | ||
| .security-review | ||
| docs | ||
| scripts | ||
| tests | ||
| .env.example | ||
| .gitignore | ||
| agents.py | ||
| conftest.py | ||
| graph.py | ||
| models.py | ||
| README.md | ||
| requirements.txt | ||
| retriever.py | ||
| run.py | ||
| state.py | ||
| telemetry.py | ||
| tools.py | ||
orchestrator — DEPRECATED
This repository is deprecated and archived as of 2026-07-14.
The agent-team subsystem this repo supported was decommissioned on 2026-06-26,
superseded by the Open SWE deployment on sh-openswe. With that subsystem gone,
the LangGraph router and its supporting agents no longer have a durable purpose.
What moved where
- Security review gate/sweep subsystem moved to
Sea-Haven-Industries/security-reviewon 2026-06-29. That repo is now the source of truth forreview.sh,nightly_sweep.sh, the checker library, and the git hooks. - Cross-family (GPT-4.1) review now runs via
cross_review.pydirectly inSea-Haven-Industries/security-review— a direct CLI with no routing layer, replacing thecross_reviewerroute through this repo'srun.py. - Automated security sweeps run as Claude Code web cloud routines rather than the R720 VM nightly sweep described in this repo's old docs.
Status
No further changes will be made here. Full commit history is preserved for
reference. Do not build on this repo — use
Sea-Haven-Industries/security-review
and the Open SWE deployment on sh-openswe instead.