majiayu000/spellbook / skills/sol-luna-router
sol-luna-router
VerifiedRoute substantial coding or repository-review work through a two-stage Sol commander/verifier and separate Luna Max CLI worker. Use when the user explicitly asks Sol to direct or supervise Luna, explicitly requests this router, needs an isolated auditable Luna-owned implementation with a Sol review loop, or asks to diagnose or measure the router itself. Do not use merely because a task involves coding, verification, progress reporting, or token saving. For small or straightforward tasks, keep work in Sol unless the user explicitly requests Luna; router configuration, usage, and efficiency questions should use read-only inspection without launching a worker.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 13.1 KB |
| agents/openai.yaml | config | 234 B |
| evals/transport-warning-benchmark-2026-08-12-task.md | markdown | 1.6 KB |
| evals/transport-warning-benchmark-2026-08-12.json | config | 5.1 KB |
| references/rate-card-2026-08-05.json | config | 548 B |
| references/run-log.md | markdown | 7.4 KB |
| references/transport-warning-benchmark-2026-08-12.md | markdown | 2.7 KB |
| scripts/analyze_run_log.py | script | 41.8 KB |
| scripts/run_luna_worker.py | script | 31.6 KB |
| scripts/test_analyze_run_log.py | script | 30.8 KB |
| scripts/test_benchmark_evidence.py | script | 38.8 KB |
| scripts/test_run_luna_worker.py | script | 35.5 KB |
| scripts/worker_events.py | script | 1.6 KB |
evals/transport-warning-benchmark-2026-08-12-task.md
View on GitHub# Task: preserve recovered transport warnings
The `sol-luna-router` runner consumes Codex JSONL events. A transport-level `error`
event or an `item.completed` event whose item type is `error` can be transient: Codex
may reconnect and later emit a valid final agent message and `turn.completed`.
Implement the following behavior:
1. A run that exits zero and ultimately contains a thread id, final agent message, and
`turn.completed` must succeed even if transient error events appeared earlier.
2. Preserve those recovered error messages, in event order, in a `warnings` list in the
parsed result.
3. `turn.failed` remains fatal even if later completion-looking events are present.
4. A nonzero process exit remains fatal and should surface the best available error.
5. A recovered error followed by an incomplete event stream or no final agent message
must still fail closed for the missing completion evidence.
6. Add focused regression tests for the new behavior without weakening existing tests.
Allowed files:
- `skills/sol-luna-router/scripts/run_luna_worker.py`
- `skills/sol-luna-router/scripts/test_run_luna_worker.py`
- `skills/sol-luna-router/SKILL.md` only if the observable contract needs documentation
Do not inspect other checkouts, external repositories, remote history, or commits beyond
this isolated repository. Do not change generated registries or unrelated files.
Verification:
```bash
python3 skills/sol-luna-router/scripts/test_run_luna_worker.py
python3 -m py_compile skills/sol-luna-router/scripts/run_luna_worker.py
```
Return the root cause, changed files, verification outcomes, and remaining risks.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| rea-analysis-optimizationmorluto/reaOptimize REA providers and shared analysis workflows for agent context, latency, CPU, and memory while preserving correctness and Evidence. Use for analysis bottlenecks and cache or query design; CI/test-suite optimization belongs to rea-ci-tests-audit. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| rea-ci-tests-auditmorluto/reaAudit REA CI and test suites for correctness, fixture drift, false passes, and measured performance; use for suite or workflow investigations and efficiency changes, rather than ordinary feature testing or a single test failure. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first. | Stars: 196k | Updated | Claude CodeCodex | Verified | Install |