majiayu000/spellbook / skills/capability-distill
capability-distill
Verified能力蒸馏工作流——从用户批准的强模型访谈或真实任务轨迹中提取非显然的判断规则,形成可审计的 judgment packet,再交给 skill-audit 和 skill-creator 决定是否落成可加载 skill。当用户说“蒸馏这个模型的判断力”“把这次任务的关键决策固化下来”“模型窗口要关了,保留它在某类场景的判断”时使用。普通流程文档、直接写 SKILL.md、泛化最佳实践或未授权的会话日志扫描不使用本 skill。
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 10.1 KB |
| evals/evals.json | config | 4.1 KB |
| evals/files/incident_existing_skills.md | markdown | 759 B |
| evals/files/incident_timeline.md | markdown | 1.4 KB |
| evals/files/review_judgment_packet_v1.md | markdown | 1.1 KB |
| evals/files/review_trajectory.md | markdown | 1.3 KB |
evals/files/incident_timeline.md
View on GitHub# Redacted incident timeline
All service, customer, host, and credential identifiers have been removed.
- 09:05 — API latency and connection wait time rose together. Candidate causes were database lock contention, pool exhaustion, or a dependency outage.
- 09:08 — Lock-wait metrics stayed flat while every application pool was saturated and request queues grew. The responder chose to inspect concurrency and pool limits instead of restarting the database.
- 09:12 — The preceding deploy had increased worker concurrency without changing pool capacity. The responder capped concurrency on one canary instance. Rollout would continue only if pool use stayed below 80% and queue depth declined for two observation windows.
- 09:18 — The canary met both signals, so the cap was rolled out. Median latency recovered and the main alert cleared.
- 09:24 — Retry volume continued rising even though latency looked normal. The responder treated the cleared alert as a false-success signal and paused closure.
- 09:28 — A client retry policy was amplifying intermittent timeouts. The responder reduced retry concurrency on a canary, requiring retry volume and error rate to fall without queue growth.
- 09:36 — Both canary checks held for two windows, the change was rolled out, and the incident was closed with follow-up work for capacity testing.
Outcome: no database restart occurred; the two canary changes were reversible and restored stable service.
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: 194k | Updated | Claude CodeCodex | Verified | Install |