eve-cli-primitives

Eve CLI 原语

使用此技能作为 Eve 的命令指南。示例应简短具体。

配置文件（API 目标 + 默认值）

# Create and use a profile
eve profile create staging --api-url https://api.eve.example.com
eve profile use staging

# Set defaults to avoid repeating flags
eve profile set --default-email [email protected] --default-ssh-key ~/.ssh/id_ed25519
eve profile set --org org_xxx --project proj_xxx

# Inspect current profile
eve profile show

身份验证

eve auth login
eve auth status
eve auth logout
eve auth permissions

# Sync local OAuth tokens for agent harnesses (optional)
eve auth sync

# Nostr challenge (API-level, no CLI command yet)
# POST /auth/challenge {"provider": "nostr", "pubkey": "<hex>"}

组织和项目

eve org list
eve org ensure my-org --slug myorg

eve project list
eve project ensure --name "My App" --slug my-app --repo-url [email protected]:me/my-app.git --branch main

# Membership management
eve org members --org org_xxx
eve org members add [email protected] --role admin --org org_xxx
eve org members remove user_abc --org org_xxx

eve project members --project proj_xxx
eve project members add [email protected] --role admin --project proj_xxx
eve project members remove user_abc --project proj_xxx

URL 影响： 组织的 --slug 和项目的 --slug 会直接构成部署 URL 和 K8s 命名空间：

• URL：{service}.{orgSlug}-{projectSlug}-{env}.{domain}（例如，api.myorg-my-app-staging.eve.example.com）
• 命名空间：eve-{orgSlug}-{projectSlug}-{env}（例如，eve-myorg-my-app-staging）

请谨慎选择 slug——创建后不可更改。

环境和部署

# Create a persistent environment
eve env create staging --project proj_xxx --type persistent

# Inspect environments
eve env list --project proj_xxx
eve env show staging --project proj_xxx

# Deploy an environment (requires --ref with 40-char SHA or a ref resolved against --repo-dir)
eve env deploy staging --ref main --repo-dir .

# When environment has a pipeline configured, the above triggers the pipeline.
# Use --direct to bypass pipeline and deploy directly:
eve env deploy staging --ref main --repo-dir . --direct

# Pass inputs to pipeline:
eve env deploy staging --ref main --repo-dir . --inputs '{"key":"value"}'

# Diagnostics
eve env diagnose proj_xxx staging
eve env diagnose proj_xxx staging --request <req_id>     # request-level integrated diagnose
eve env logs proj_xxx staging <service>
eve env logs proj_xxx staging <service> --follow         # stream logs
eve env logs proj_xxx staging <service> --filter route=/api/x --filter status=500
eve env delete proj_xxx staging

--filter k=v 可重复使用；常见键包括：route、status、level、request_id。使用 --grep 进行自由文本匹配，使用 --since/--tail 限定时间范围。

作业（创建 + 观察）

eve job create --description "Review auth flow"
eve job list --phase active
eve job show <job-id>
eve job follow <job-id>
eve job diagnose <job-id>
eve job result <job-id>

# Per-job harness + env overrides (Eden-style per-request brain selection)
eve job create --description "..." \
  --harness-override-file ./profile.json \
  --env-override OPENAI_API_KEY='${secret.OPENAI_KEY}' \
  --env-override MODEL_ENDPOINT=https://qwen.local

--harness-override-file 接受 JSON 对象 {harness, model?, reasoning_effort?, variant?, temperature?}。--env-override KEY=VALUE 可重复使用，会验证 UPPER_SNAKE_CASE 键，并支持 ${secret.KEY} 插值，该插值在启动时解析。

协调（监督 + 线程）

eve supervise <job-id> --timeout 60
eve thread messages <thread-id> --since 10m
eve thread post <thread-id> --body '{"kind":"directive","body":"focus on auth"}'
eve thread follow <thread-id>

Agent + 聊天

# Sync agent/team/chat configuration from repo
eve agents sync --project proj_xxx --ref main --repo-dir .
# Inspect resolved agent config (from latest sync)
eve agents config --project proj_xxx --json
# Simulate chat routing without Slack
eve chat simulate --project proj_xxx --team-id T123 --channel-id C456 --user-id U789 --text "hello"

# Set org default agent slug (fallback when no slug matches)
eve org update org_xxx --default-agent mission-control

包（AgentPacks）

eve packs status
eve packs resolve --dry-run
eve migrate skills-to-packs

Slack 网关命令（在 Slack 内运行）：

@eve <agent-slug> <command>
@eve agents list
@eve agents listen <agent-slug>
@eve agents unlisten <agent-slug>
@eve agents listening

Nostr 命令（通过私信发送至平台公钥）：

/agent-slug command text
agent-slug: command text

密钥

eve secrets list --project proj_xxx
eve secrets set API_KEY "value" --project proj_xxx
eve secrets delete API_KEY --project proj_xxx
eve secrets show API_KEY --project proj_xxx
eve secrets ensure --project proj_xxx --keys API_KEY
eve secrets export --project proj_xxx --keys API_KEY

系统管理（如已获授权）

eve system status
eve system jobs
eve system envs
eve system env-health --json
eve system env-cost --all
eve system logs api --tail 50
eve system pods
eve system events

eve system env-cost 从管理 API 返回本月至今的环境分配快照。这是快照视图，而非实时 OpenCost 查询；输出中会标注过时或不可用的估算值。

工作流

eve workflow list
eve workflow run qa-review --input '{"task":"audit"}'
eve workflow invoke qa-review --input '{"task":"audit"}'
eve workflow run qa-review --input '{...}' --env-override MODEL=opus-4.5
eve workflow logs job_abc123

--env-override KEY=VALUE 可在 run 和 invoke 中重复使用；这些值会作为 env_overrides 传递给每个步骤的作业。

集成（Slack、Nostr、GitHub）

eve integrations list --org org_xxx
eve integrations slack connect --org org_xxx --team-id T123 --token xoxb-test
eve integrations test <integration_id> --org org_xxx

# Nostr integration (via API)
# POST /integrations {"provider": "nostr", "account_id": "<platform-pubkey>", ...}

流水线和工作流

eve pipeline list
eve pipeline show <project> <name>
eve pipeline run <name> --ref <sha> --env <env> --repo-dir ./my-app

eve workflow list
eve workflow show <project> <name>
eve workflow run <project> <name> --input '{"k":"v"}' [--env-override KEY=VALUE]
eve workflow invoke <project> <name> --input '{"k":"v"}' [--env-override KEY=VALUE] [--no-wait]
eve workflow logs <job-id>

构建

构建是一等原语，用于跟踪从输入（规范）到执行（运行）再到输出（制品）的镜像构建过程。

# List builds for a project
eve build list [--project <id>]

# Create a build spec
eve build create --project <id> --ref <sha> --manifest-hash <hash>

# Show build spec details
eve build show <build_id>

# Start a build run
eve build run <build_id>

# List runs for a build
eve build runs <build_id>

# View build logs
eve build logs <build_id> [--run <run_id>]

# List produced image artifacts (digests)
eve build artifacts <build_id>

# Full diagnostic dump (spec + runs + artifacts + logs)
eve build diagnose <build_id>

# Cancel an active build
eve build cancel <build_id>

流水线的 build 步骤会自动执行构建。使用 eve build diagnose 调试构建失败。

跟踪（OTEL）

可供 Agent 查询的跟踪存储。结合使用 eve env diagnose --request <id> 和 eve env logs --filter request_id=<id>，进行端到端请求排查。

eve traces query --request-id <id>                    # spans for a single request
eve traces query --trace-id <id>                      # spans for a known trace
eve traces query --service api --since 15m --error    # recent errors in a service
eve traces query --route /api/projects --p99          # p99 latency for a route
eve traces query --request-id <id> --json             # machine-readable

必须指定 --request-id、--trace-id、--since、--route 或 --error 之一。--no-cache 会绕过跟踪缓存。

自定义域名

eve domain list [--env <name>] [--project <id>]
eve domain verify <hostname>             # check DNS and show activation status
eve domain status <hostname>             # show domain status and owning env
eve domain transfer <hostname> --to <env>  # move ownership between envs
eve domain unbind <hostname>             # clear binding so next deploy claims it
eve domain remove <hostname>             # delete the domain entirely

系统运行状况

eve system health

Harness（可选）

eve harness list
eve harness list --capabilities
eve harness get mclaude
eve agents config --json

注意事项

• 如果配置文件中未设置默认项目，请使用 --project。
• 集成按组织划分；集成相关命令请使用 --org。
• 作业相关命令优先使用 eve job ...（单数形式）。