feat(agent): unify models and validate OpenCode providers

Generate chat, task and learning models from agent-models.json.
Use WaterNetwork fast/expert models with 272k context and validate
upstream names before readiness. Add an isolated model check command
with signal-safe cleanup and plugin, Skill and MCP management docs.

Verified in isolation: type checks, 26 focused tests and both model
catalog checks; no model inference requests.
This commit is contained in:
2026-10-09 11:28:37 +08:00
parent 75d8d526f9
commit d58bd19039
30 changed files with 1125 additions and 469 deletions
+7 -3
View File
@@ -12,9 +12,13 @@ OPENCODE_MODE=embedded
OPENCODE_COMMAND=opencode2
# client 模式连接私有 V2 server 时,填写其 server password(不要提交实际值)。
# OPENCODE_CLIENT_PASSWORD=
OPENCODE_MODEL=deepseek/deepseek-flash
# 可选:后台工况任务首选模型,必须同时出现在 OPENCODE_MODEL_OPTIONS 中。
# OPERATIONAL_TASK_MODEL=deepseek/deepseek-v4-pro
# 模型、provider、默认聊天/后台/学习模型统一在 agent-models.json 配置。
# 可选:使用独立的本地配置文件,修改后重启 Agent 和外部 OpenCode。
# AGENT_MODEL_CONFIG_PATH=./agent-models.local.json
# API key 仅在本地或部署环境填写,不提交实际值。
# 当前配置未启用 DeepSeek;重新配置该 provider 时再填写。
# DEEPSEEK_API_KEY=
WATERNETWORK_API_KEY=
OPENCODE_HOSTNAME=127.0.0.1
OPENCODE_PORT=0
# client 模式必须在 Agent 与外部 OpenCode 进程配置同一个高熵值;embedded 模式可留空。