Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
git clone --depth 1 https://github.com/navraj007in/architecture-cowork-pluginWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/commands/navraj007in/architecture-cowork-plugin/agent-spec)<a href="https://agentmods.dev/commands/navraj007in/architecture-cowork-plugin/agent-spec"><img src="https://agentmods.dev/badge/commands/navraj007in/architecture-cowork-plugin/agent-spec/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/commands/navraj007in/architecture-cowork-plugin/agent-spec"><img src="https://agentmods.dev/badge/commands/navraj007in/architecture-cowork-plugin/agent-spec.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00014 | $0.01322 |
| Opus 5 | $0.00007 | $0.00661 |
| Sonnet 5 | $0.00003 | $0.00264 |
| Haiku 4.5 | $0.00001 | $0.00132 |
Grade A, and why
agent-spec scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 11d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 183 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/architect:agent-spec
Trigger
/architect:agent-spec [describe what the agent should do]
Purpose
Design a complete AI agent architecture. This is the differentiator command — most architecture tools don't cover AI agents. Produces everything a developer needs to build the agent.
Workflow
Step 1: Understand the Agent
First, check architecture-output/_state.json. If it exists, read it in full — it provides instant access to project, tech_stack, components, design, entities, and personas without reading larger files. Use its values directly where available; fall back to SDL (check solution.sdl.yaml first; if absent, read sdl/README.md then the relevant module files) only for detail not in _state.json.
If a description is provided, extract:
- What the agent should do (purpose)
- Who interacts with it (user type)
- Where it lives (chat UI, Slack bot, API, etc.)
If not enough context, ask:
"What should this agent do? Tell me: (1) what task it handles, (2) who uses it, and (3) how they interact with it (chat, Slack, API, etc.)"
Step 2: Design the Agent
Using the agent-architecture skill, determine:
- Best orchestration pattern for this use case
- Required tools
- Optimal LLM provider and model
- Memory strategy
- Guardrails needed
Step 3: Generate Output
Agent Overview
| Field | Value |
|---|---|
| Purpose | What the agent does (one sentence) |
| Interface | How users interact (chat-ui, slack-bot, api, etc.) |
| Orchestration | Pattern used (ReAct, multi-agent, etc.) |
| LLM Provider | Recommended provider and model |
| Memory | Strategy (session, persistent, vector-store) |
Why This Architecture
2-3 sentences explaining why this orchestration pattern and LLM were chosen. Use the founder-communication skill — explain in plain English, then add technical rationale.
LLM Provider Recommendation
| Criteria | Assessment |
|---|---|
| Why this provider | Rationale for choosing this LLM |
| Why this model | Rationale for the specific model tier |
| Alternative | Second-best option and when to switch |
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 11d ago First seen · 183 lines · 14 tokens per session scan A 4672c1f2d602
agent-spec is a command published in the GitHub repository navraj007in/architecture-cowork-plugin (2 stars, last pushed 2mo ago), licensed Apache-2.0. It adds 14 tokens to every session and 1,322 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other commands, from other repositories
nyann:retrofit
Audit an existing repo against a profile and fix what's drifted. Unlike doctor (read-only), retrofit detects missing hooks, misconfigured gitignore, documentation gaps, and non-compliant history, then offers to remediate via bootstrap. Idempotent — safe to re-run.
nyann:apply
Apply an Infrastructure-as-Code change — the highest-stakes mutator in nyann; it can change real cloud infrastructure. Re-runs the plan, shows it, confirms, then applies. Unmistakably opt-in: apply is never the default and destructive applies require a second explicit confirm. For IaC apply intent only (not "apply a…
nyann:hotfix
Create the branch topology for a patch release against a previously tagged version. Ensures release/ . exists from the source tag, then creates hotfix/ off it. After this, the user commits the fix and runs /nyann:release from the hotfix branch.
nyann:release
Cut a release: generate a CHANGELOG section from Conventional Commits, make a release commit, and create an annotated tag. Defaults to conventional-changelog strategy.
nyann:ship
Open a GitHub pull request AND merge it in one step. Default uses GitHub's native auto-merge (returns immediately with outcome:"queued"); --client-side polls for green CI in the foreground then runs gh pr merge. Requires gh installed + authed.
nyann:cleanup-branches
Prune local branches whose work is already merged into the base. Lists candidates first, then applies on --yes. Mirrors the safe-delete semantics of git branch -d (lowercase d): nothing unmerged is touched.