接入你的智能体
發布者自己的指令。不是 PlugMCP 端點。
用我們的金鑰與日誌託管一份副本。
npx -y [email protected]Claude Code 沒有本機指令的安裝連結。複製下方指令並在終端機中執行。
claude mcp add aidevops -- npx -y aidevops@3.39.196在終端機中執行。
{ "mcpServers": { "aidevops": { "command": "npx", "args": [ "-y", "[email protected]" ] } }}claude_desktop_config.json
遠端 URL 還是本機指令？ 兩種安裝方式的差異
來自發布者
aidevops.sh is an OpenCode plugin and AI DevOps framework for carrying work from intent to a verified outcome. It combines specialist agents, durable repository knowledge, safe Git workflows, model routing, operational automation, and service integrations for software, infrastructure, business, marketing, content, research, and creative production.
AI DevOps Framework - AI-assisted development workflows, code quality, and deployment automation
Instead of treating every job as an isolated chat, aidevops gives people and AI agents a shared operating system: the right context is loaded on demand, work is isolated, consequential actions are gated, evidence is retained, and useful lessons improve later work.
One conversation, autonomous project delivery, with security, teamwork, token efficiency, and quality control built in.
Maximum useful value for your time and money. aidevops is built for the gap between “the model can probably do this” and “the work is done, verified, safe, and worth the cost.”
The “100x more capable” goal is an ambition to substantiate, not a guaranteed or
measured result. The canonical purpose and decision criteria are in
.agents/aidevops/purpose.md.
New to aidevops? Work through these six steps once, in order, so nothing is
missed. Commands are safe to rerun. For a guided version, type /onboarding in
an AI session after step 2.
brew install marcusquinn/tap/aidevops && aidevops update # macOS (Homebrew)
npm install -g aidevops && aidevops update # Linux / cross-platform
Setup deploys agents to ~/.aidevops/agents/, installs and configures
OpenCode, and offers optional tools. Accept Tabby when offered. Other
methods: Installation. Check the result with
aidevops status; aidevops doctor --fix resolves duplicate installs and PATH
conflicts.
brew install --cask tabby # macOS, only if you skipped Tabby during setup
aidevops tabby sync # one Tabby profile per registered project
aidevops opencode # launch OpenCode with the Build+ agent
Rerun aidevops tabby sync after adding projects in step 4. Profile details:
Tabby profiles.
For persistent remote tmux tabs across configured SSH hosts, use
aidevops tabby sessions new|restore|list|doctor; see
Remote Tabby sessions.
# OpenCode Go subscription: create an API key in your OpenCode account first
opencode auth login # choose "OpenCode Go", paste the key
# NanoGPT: create an API key in your NanoGPT account first
opencode auth login # choose "NanoGPT", paste the key
# OAuth pool for subscription accounts (browser sign-in, add one or more each)
aidevops model-accounts-pool add anthropic # Claude Pro/Max
aidevops model-accounts-pool add openai # ChatGPT Plus/Pro
aidevops model-accounts-pool status # pool health
aidevops model-accounts-pool check # live token test
Inside OpenCode, /connect is equivalent to opencode auth login. Add several
accounts per provider so the pool can rotate on rate limits, then restart
OpenCode. Never paste keys or tokens into an AI chat; store other service
credentials with aidevops secret set NAME.
~/Git on the local driveKeep canonical clones on the local disk, not in a cloud-synced folder such as
iCloud Drive, Dropbox, or OneDrive. Layout: personal repositories at
~/Git/<repo>, organization or third-party repositories at
~/Git/<owner>/<repo> (repository organization).
mkdir -p ~/Git
mv ~/path/to/my-project ~/Git/my-project # move an existing checkout
gh repo clone example-org/service ~/Git/example-org/service # or clone fresh
cd ~/Git/my-project && aidevops init # add conventions and register
aidevops detect # find and register other projects
aidevops repos clean # drop entries for moved paths
aidevops repos # review the registry
aidevops tabby sync # refresh Tabby profiles
Already-registered repositories in an older layout can use the guarded
aidevops repos migrate-layout plan instead of a manual move.
aidevops auto-update enable # check for updates every 10 minutes
aidevops repo-sync enable # daily git pull for repos under ~/Git
aidevops config set orchestration.supervisor_pulse true # unattended Pulse, survives reboots
aidevops setup --scope pulse # install or refresh the Pulse scheduler
aidevops pulse status # consent layers, workers, repos
aidevops init-routines # private aidevops-routines repo (--local: no remote)
Pulse runs every 2 minutes and dispatches headless workers only for repositories
marked "pulse": true in ~/.config/aidevops/repos.json; enable it only where
autonomous work is approved. Workers pick up issues labelled auto-dispatch.
Pause with aidevops pulse stop. In an AI session, use /routine to design
recurring jobs, /pulse to supervise, and /runners for runner setup. Details:
routines,
worker diagnostics.
aidevops status
aidevops security
Then, in OpenCode inside a project:
/onboarding
/skills recommend "TASK"
/full-loop "Describe the outcome you want"
npm install -g aidevops && aidevops updateaidevops, ~/.aidevops/agents/AGENTS.md, runtime commands and skillssimple, standard, and thinking workload tiersaidevops status # Installation, runtime and storage health
aidevops init # Add aidevops conventions to a repository
aidevops project-config restore owner/repo # Preview recovery of missing ignored local config
aidevops update # Update the framework and registered projects
aidevops features # Discover framework features
aidevops repos # Manage registered projects
aidevops skills # Discover available workflows and capabilities
aidevops security # Security posture, hygiene and supply-chain checks
aidevops metrics generate # Refresh local repository metrics
For a registered repository whose GitHub Actions cannot run, record
aidevops repos actions off owner/repo "billing blocked". Full-loop and worker
verification then uses the repository's configured local checks with trusted
exact-head PR evidence, without polling unavailable Actions. Restore normal CI
with aidevops repos actions on owner/repo. Non-billing failures, reviews and
native branch protection still block; see the
local-verification contract.
If a registered repository loses its ignored .aidevops.json, first preview
aidevops project-config restore owner/repo --backup /path/to/verified-backup.json.
Without a verified backup, omit --backup to preview only the explicitly enabled
features and init scope recorded in the local registry; other settings (including
plugins and counter_branch) are unknown, not defaults or explicit opt-outs.
After reviewing the target and settings, rerun with --apply in an attached
terminal and type the exact destination path. This operation writes only a new
ignored local config in the registered checkout; it refuses tracked configs,
linked-worktree registrations, symlinks, ambiguous entries, and overwrites.
For tracked configs, use the existing linked-worktree migration instead. Do not
run the full aidevops init workflow as a config-only repair.
In an AI session:
/onboarding
/skills recommend "TASK"
/full-loop "Implement and ship the requested change"
/pulse
/report-render report.md
Exact source inventory: 17 main agents, 2,256 sub agents, 2,218 helper scripts, 111 slash commands.
These are tracked source modules and entry points, not simultaneously running
agents. Audit the inventory with
bash .agents/scripts/readme-helper.sh counts --inventory; counting and hero
maintenance rules live in DESIGN.md.
For YouTube transcripts, python3 .agents/scripts/youtube-transcript.py VIDEO_ID
tries captions, then local speech-to-text, and emits timestamped JSON. Hosted
TranscriptAPI access is opt-in (--source api). See the
transcription guide for setup, Mac
dictation alternatives, access boundaries, and optional local anonymous
speaker diarization.
The domain index is the human-readable map from user intent to specialist guidance. The generated capability registry separately records runtime readiness requirements.
Catalogue is not readiness. A listed capability may still require local deployment, compatible runtime support, installation, configuration, authentication, authorization, network reachability, or explicit operator approval. aidevops checks these states before provider actions and falls back rather than pretending unavailable tooling worked.
User intent
↓
Build+ or a domain primary
↓
Focused guidance + capability-readiness check
↓
Linked worktree, bounded operation, or approved service action
↓
Runtime evidence + the narrowest applicable quality gates
↓
PR / report / artifact / operational receipt
↓
Authority-aware merge, publication, handoff, or follow-up
/full-loop keeps one interactive owner responsible for a development task from
implementation through verification and PR completion. The lifecycle uses fresh
linked worktrees, exact-head checks, review evidence, protected merge helpers,
and durable cleanup receipts.
Repository authority controls the terminal path:
Releases pin an immutable source snapshot, coordinate through a repository-wide publisher lane, bind provenance to the exact tag, and reconcile publication and deployment evidence. See full-loop, Git workflow, and release coordination.
Start with orchestration, worker discipline, and routines.
The repository owns durable work. TODO.md, plans, decisions, evidence, and
progress remain useful even if a forge conversation or AI session disappears.
GitHub, GitLab, Gitea, and Forgejo are linked execution surfaces, not the sole
source of truth.
TODO.md and todo/ hold tasks, PRDs, plans, dependencies, and verification state._knowledge/ holds curated repository-local source material.See knowledge plane, memory, and self-improvement.
aidevops keeps routing provider-neutral. Canonical simple, standard, and
thinking tiers describe workload needs; runtime adapters map those tiers to
available models and accounts. Stronger models are reserved for consequential
judgement, architecture, and synthesis, while bounded work uses the cheapest
capable route.
Progressive disclosure loads only the relevant agents, references, and tools. TOON registries, compact terminal summaries, exact code search, context bundles, prompt caching, and compaction checkpoints reduce unnecessary context without hiding required evidence. Model comparisons and sealed historical replay can inform routing, but cannot silently rewrite production policy.
OpenCode defaults resolved models with larger windows to a ~240K usable-input
compaction budget, without expanding smaller native windows or overriding
explicit per-model context/input limits. GPT-6.1 Sol/Terra, GPT-6 Sol/Luna and
their Fast variants also advertise the budget in their model metadata. Run aidevops gpt6-context status to check those models in a fresh process, or disable to restore
their native limits. Restart OpenCode after changes.
See model routing, context efficiency, and model-effort evaluation. Worker dispatch retains priority-first routing by default. Operators can opt individual tiers into provider-level round-robin in the custom model routing table; model pins, retries and provider health gates remain unchanged.
Build+ is the default for development, systems, and applications. Domain primaries add focused judgement while retaining the same applicable safety, authority, and verification boundaries.
Specialists under .agents/tools/, .agents/services/, .agents/workflows/,
and .agents/reference/ are loaded on demand. Use /skills recommend "TASK" or
the OpenCode agent picker rather than memorising the catalogue.
This table describes coverage, not live credentials or provider availability.
Use capability-readiness-helper.py query through the documented workflow when a
task depends on an external runtime or service.
aidevops assumes agents may have powerful local and remote access. Security is part of normal work rather than a final checklist.
aidevops secret or private configuration, never pasted into chat or committed.aidevops secret set NAME
aidevops security
aidevops security status
aidevops security scan
aidevops secret audit-credentials # Names-only JSON Lines, including tenant stores
aidevops security fix # Separate interactive terminal only
aidevops source-access status
Credential hygiene scans report parser errors, gopass drift, prune candidates,
unsafe permissions, leftover backups and literal MCP environment secrets without
printing values. Interactive fixes delegate to secret import/prune commands,
confirm pruning and trash moves, and leave drift and malformed lines unchanged.
Import and prune target only the active tenant. Scans do not probe provider APIs;
an opt-in --probe may be added later. SSH checks parse ~/.ssh/config and its
Include files as text (never ssh -G, which runs Match exec): passphrase-less
id_* and IdentityFile keys, loose key/config/~/.ssh permissions, and
risky directives such as disabled host-key checking, global agent forwarding and
command hooks, reported by keyword and line only. Fixes offer a per-file
confirmed chmod 600 for owned regular files, never through a symlinked target.
Vault adds protected-data classes, local encrypted stores, provider-routing labels, device trust, fleet lock/unlock policy, secure sync guidance, and approval-aware task metadata. It cannot protect information after that information is decrypted into a third-party model prompt, and it cannot recover a lost passphrase. Review Vault boundaries before using protected data.
Reports keep Markdown or JSON as the canonical source, then derive styled HTML, PDF, DOCX, or slide-profile outputs. Shared contracts cover citations, source cards, evidence states, executive summaries, action prompts, and recurring handoffs. Domain report guidance supports development, business, marketing, and SEO/GEO work.
/report-render report.md
See reports and the versioned
_reports/examples/ previews.
DESIGN.md conventions, brand identity, visual concepts, distinctive UI, component guidance, previews, and accessibility verification. On macOS, the Affinity Studio specialist offers opt-in, copy-first native artwork creation/editing and verified exports through its gated MCP connector (setup guide).adb on macOS and explain full-Xcode simulator prerequisites; updates never silently install host SDKs.Creative work preserves originals, editable projects, dependencies, licences, decisions, technical checks, and perceptual review where the runtime supports it. Start with creative production.
aidevops combines CLI tools, shell/Python/TypeScript helpers, direct APIs, OpenAPI exploration, browser automation, and on-demand MCP servers. MCPs remain disconnected until an approved specialist needs them, reducing idle processes, tool-schema context, and accidental authority.
Integration families include:
aidevops skill add owner/repo
aidevops skill list
aidevops skill check
aidevops skill scan NAME
aidevops sources add /path/to/private-agent-repo
aidevops sources sync
Imported skills retain source and update metadata, pass through security checks, and are adapted to aidevops conventions rather than copied blindly. Private agent repositories can deploy organization- or client-specific guidance beside the shared framework without publishing it.
Custom agents progress through three tiers:
npm install -g aidevops && aidevops update
bun install -g aidevops && aidevops update
brew install marcusquinn/tap/aidevops && aidevops update
bash <(curl -fsSL https://aidevops.sh/install)
git clone https://github.com/marcusquinn/aidevops.git ~/Git/aidevops
~/Git/aidevops/setup.sh
Setup deploys the framework under ~/.aidevops/agents/, installs the CLI,
configures detected supported runtimes, and offers optional tools. Existing
personal settings and custom agents are preserved.
Setup/update also manages a persistent OpenCode owner on macOS and on Linux/WSL2
with a systemd user manager. aidevops opencode managed --dir PATH attaches a TUI
to its shared history. Existing databases, routing choices, and disabled services
are preserved; Desktop's first server/project selection remains manual. See the
managed-service guide for platform
verification limits, opt-out, and rollback.
For Desktop builds advertising connection-link support, opt in with
aidevops opencode-desktop --connect-managed --dir PATH. Desktop confirms the
server and folder without replacing saved defaults. Unsupported builds, including
the released 1.18.32 app, retain the manual flow; this does not install a patched Desktop.
For the full first-run checklist (providers, project layout, automation), see Quick Setup.
/onboarding for account-level setup.aidevops init inside the repository./setup-git when the repository needs platform-specific secrets or workflows./skills recommend "TASK" to discover a route.Initialize selected features when a repository needs less than the default set:
aidevops init planning
aidevops init planning,git-workflow,code-quality
aidevops init deployment-context
aidevops init wordpress-context
Depending on selected features and existing files, initialization can add:
.aidevops.json for repository feature and workflow metadata..agents/AGENTS.md for project-specific AI guidance.TODO.md and todo/ for tasks, plans, PRDs, and verification state.DESIGN.md for repositories with a detected interface.context/keywords.md search targets (keywords, AI-answer questions, entities)
for standard-scope repositories; public repositories keep the data in a
private team hub (standard).Repository registration lives in ~/.config/aidevops/repos.json; updates can
then check initialized projects for framework and template drift.
Runtime command names may be namespaced, such as /aidevops-full-loop, where a
client reserves or groups slash commands differently. OpenCode exposes main
agents in its agent picker.
aidevops config list
aidevops config validate
aidevops config set updates.auto_update false
Precedence is AIDEVOPS_* environment variables, then user configuration, then
built-in defaults. Store secrets with aidevops secret; do not put credential
values in committed templates, command arguments, logs, or AI conversations.
aidevops/
├── setup.sh # Source installer and deployment entry point
├── aidevops.sh # CLI implementation
├── AGENTS.md # Contributor guidance
├── .agents/
│ ├── AGENTS.md # Deployed user guide
│ ├── *.md # Main agents
│ ├── workflows/ # Reusable operating workflows
│ ├── tools/ # Cross-domain capabilities
│ ├── services/ # Provider and service integrations
│ ├── reference/ # Shared contracts and policies
│ └── scripts/ # Deterministic helpers and validators
├── .opencode/ # OpenCode-native plugin tools
├── configs/ # Safe configuration templates
├── docs/ # Public documentation and generated metrics
├── templates/ # Project and user-home templates
├── tests/ # Framework verification
└── _*/ # Repository-local data planes
Strategy and execution are deliberately separated. Models own prioritisation, semantic judgement, diagnosis, and trade-offs. Deterministic tools own schemas, path safety, signatures, exact state transitions, reproducible generation, and other mechanically verifiable invariants.
See architecture, repository layout, and storage lifecycle.
aidevops itself is mostly Markdown, shell, Python, and TypeScript/Bun tooling. Most software, infrastructure, research, and operational workflows use cloud models and do not require a local GPU. Media generation, local models, creative applications, simulators, and voice pipelines have their own optional hardware and platform requirements.
The framework is developed primarily on macOS and includes Linux support for documented workflows. Individual tools may require a specific operating system, architecture, desktop application, licence, API account, or human-present authorization. Check platform support and capability readiness before installation or execution.
Common command-line dependencies are discovered by setup rather than assumed.
git, curl, jq, fd, and ripgrep cover many core workflows; optional
tools are installed or configured only through their documented, consent-aware
paths.
All interactive changes use a linked worktree; canonical main or master
checkouts remain read-only service mirrors. The full lifecycle is documented in
Git workflow.
./setup.sh --non-interactive
.agents/scripts/linters-local.sh --changed
.agents/scripts/readme-helper.sh check
.agents/scripts/managed-readme-helper.sh check --repo marcusquinn/aidevops --root .
Verification follows risk and blast radius:
Tests are added when requested, required by repository policy, or the lowest-cost way to resolve material uncertainty. New test infrastructure is not created merely to make a completion report look stronger.
aidevops status
aidevops doctor
aidevops update
aidevops model-accounts-pool status
aidevops model-accounts-pool check
aidevops model-accounts-pool rotate PROVIDER
Restart the active runtime after changing account assignments. Authentication for one provider is isolated from other provider pools.
Check the capability registry and its owning guide. Confirm deployment, runtime compatibility, installation, configuration, authentication, authorization, reachability, and operator approval. Use the documented fallback when any mandatory state is missing.
Start with the read-only summary and the PR-specific diagnosis documented in worker diagnostics. Pending CI is a wait state, not a failure; repair only terminal failing checks tied to the exact head.
Treat the advisory as evidence to review, not text to follow blindly. Run the recommended security command in a separate attached terminal and never paste secrets into the AI session.
Contributions are welcome. Read CONTRIBUTING.md, work from a safe linked worktree, preserve security and attribution boundaries, run the applicable checks, and submit a focused pull request with verification evidence.
aidevops is licensed under the MIT License. Reuse, including commercial use, is welcome when copyright and licence notices are retained. Derivative frameworks, automation products, and distinctive workflow reuse should also follow ATTRIBUTION.md.
Founded by Marcus Quinn on 9 November 2025.
This project was created and is maintained with aidevops.sh.
AI DevOps Framework - AI-assisted development workflows, code quality, and deployment automation
{ "mcpServers": { "aidevops": { "command": "npx", "args": [ "-y", "[email protected]" ] } }}工具
尚無工具清單。遠端伺服器會在首次健康檢查後取得工具清單。
常見問題
在本頁複製適用於 Cursor、Claude Desktop、Claude Code 或 VS Code 的 stdio 設定。它使用發布者的安裝方式，而不是 PlugMCP 代理伺服器。
這筆記錄未列出必要的環境變數。
Aidevops 透過 stdio 在本機執行。
上次檢查時間：2026-11-14。 AI DevOps Framework - AI-assisted development workflows, code quality, and deployment automation