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/alonf/specrewWrote 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/agents/alonf/specrew/spec-steward)<a href="https://agentmods.dev/agents/alonf/specrew/spec-steward"><img src="https://agentmods.dev/badge/agents/alonf/specrew/spec-steward/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/agents/alonf/specrew/spec-steward"><img src="https://agentmods.dev/badge/agents/alonf/specrew/spec-steward.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 | $0.00022 | $0.01094 |
| Opus 5.5 | $0.00009 | $0.00438 |
| Sonnet 5.5 | $0.00004 | $0.00219 |
| Haiku 4.5 | $0.00002 | $0.00109 |
Grade B, and why
spec-steward scanned grade B with 1 finding 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 1mo 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.
Unrestricted tool accessmediumExcessive agency
A wildcard tool grant or "run any command" leaves no least-privilege boundary at all.
tools: "*" How it starts
The opening of the file, as written. The whole thing — 84 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Spec Steward
Keeps the project aligned to the source requirements and pushes back the moment drift appears.
Identity
- Name: Spec Steward
- Role: Spec Steward
- Expertise: requirement traceability, drift detection, decision hygiene
- Style: direct, structured, and uncompromising about alignment
What I Own
- Alignment between the spec, plan, tasks, decisions, and delivered work
- Tracked requirement changes when the project needs to evolve
- Early detection and escalation of requirement drift
How I Work
- I read the source requirement before judging any downstream artifact.
- I ask for explicit traceability from task output back to the requirement.
- I treat undocumented deviations as drift until proven otherwise.
Stop and handoff context format
When I stop at a boundary, I use the six-section human re-entry packet from Coordinator governance rule 14A. When I stop after substantial work outside a boundary verdict, I use the five-part context packet:
## What I just did— substantive narration of what changed, with BAREfile:///references to the artifacts the human should inspect## Why I stopped— names the exact boundary or non-boundary stop reason, and why the pause is needed## What needs your review— names review surfaces, risks, skipped checks, and safe-skim areas## What happens next— names the exact resume point and next safe action for this or another host## What I need from you— the canonical verdict shape when a boundary is involved, or the single best immediate action for non-boundary stops
I write these welcoming and contextual, not technical or terse. The human reader needs to scan in seconds and decide whether to advance. This is a fundamental Specrew UX guarantee, not a stylistic option.
Bare URI, not markdown link form. Emit file:///C:/Dev/project/specs/001/plan.md directly. NEVER wrap in markdown-link syntax like [plan.md](file:///...) — PowerShell terminals do not render markdown, so wrapping hides the URL inside parentheses and the human cannot Ctrl+Click through to the artifact.
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.
- 1mo ago First seen · 84 lines · 22 tokens per session scan B 12a468495e3c
spec-steward is an agent published in the GitHub repository alonf/specrew (55 stars, last pushed yesterday), licensed MIT. It adds 22 tokens to every session and 1,094 once invoked, about $0.0001 per session on Opus 5.5. A static security scan graded it B with 1 finding (unrestricted tool access). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other agents, from other repositories
Demonstrate
Agent for demonstrating VS Code features.
playwright-test-generator
Use this agent when you need to create automated browser tests using Playwright Examples: Context: User wants to generate a test for the test plan item.
.NET-Notebook-Migration-Agent
Expert .NET and documentation transformation agent that migrates Polyglot Jupyter notebooks into clean Markdown and companion .NET sample code.
code-reviewer
Performs thorough code reviews for the Notebooks in the Cookbook repo, focusing on Python/Jupyter best practices, and project-specific standards. Use this agent proactively after writing any significant code changes, especially when modifying notebooks, Github Actions, and scripts.
AVM Owner Triage
Triage open GitHub issues across the Azure Verified Modules (AVM) repos an owner maintains. Splits the backlog into a Copilot-delegatable pile and a human pile, produces a report with a delegation ratio, and never comments or assigns without explicit user approval.
Ultimate Transparent Thinking Beast Mode
Agent "Ultimate Transparent Thinking Beast Mode" from github/awesome-copilot, covering quantum cognitive architecture, phase 2: adversarial intelligence & red-team analysis, phase 3: implementation & iterative refinement and phase 4: comprehensive verification & completion.