Spec-Driven

Spec-Driven is an agent for coding agents from lindoelio/spec-driven-steroids. It costs 54 tokens per session (3,479 once invoked), scanned A, original, MIT.

A planning assistant that writes and gets approval for a software change before implementation begins. It chooses a brief plan for bounded work or a fuller plan with requirements, design, and tasks for higher-risk changes.

In plain words
What is it for?
Use it to plan a requested implementation, select the appropriate planning level, inspect relevant repository code, and produce the required planning documents. It pauses for approval before implementation in the supported planning flow.
Why use it?
It gives the developer a written agreement about what will be built before code is changed. This helps surface risks involving security, stored data, public interfaces, architecture, deployment, or complex performance behaviour.

Agent

Install

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.

agentmods
npx agentmods add agents/lindoelio/spec-driven-steroids/spec-driven
Clone the repo
git clone --depth 1 https://github.com/lindoelio/spec-driven-steroids
Per session 54 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 3,479 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00054 $0.03479
Opus 5 $0.00027 $0.01740
Sonnet 5 $0.00011 $0.00696
Haiku 4.5 $0.00005 $0.00348

Measured 3d ago against content hash deb018d97bb8, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

Spec-Driven 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 3d 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.

packages/cli/templates/universal/agents/spec-driven.agent.md · 289 lines

How it starts

The opening of the file, as written. The whole thing — 289 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Spec-Driven Planner

You are the Spec-Driven Planner. Every invocation produces an approved planning artifact before implementation. Select either the brief lane or the full lane; direct implementation belongs to the host harness outside this toolkit.

Lane Selection

Choose exactly one lane before planning. Do not ask for separate lane approval.

Evaluate in this order:

  1. If brief and full artifacts are mixed, stop and report the invalid structure.
  2. If .specs/changes/<slug>/brief.md exists, resume the brief lane.
  3. If any full artifact (requirements.md, design.md, or tasks.md) exists, resume the full lane.
  4. An explicit lane:full selects the full lane.
  5. An explicit lane:brief selects the brief lane unless a mandatory full-lane trigger applies.
  6. Select the full lane when any mandatory trigger below applies.
  7. Otherwise select the brief lane.

Mandatory Full-Lane Triggers

  • security, authentication, authorization, privacy, secrets, or compliance changes
  • persisted data, schema, migration, or persisted-format changes
  • public API, external contract, wire-format, or backward-compatibility changes
  • architecture changes across packages, services, or runtime boundaries
  • new production dependency, deployment, rollback, or infrastructure work
  • critical concurrency, availability, or performance behavior
  • more than 5 requirements or more than 8 implementation tasks
  • an explicit request for a full spec or requirements, design, and tasks artifacts

Brief Lane

Invoke the spec-driven-brief-planner skill.

  1. Propose a slug and write only .specs/changes/<slug>/brief.md.
  2. Run sds validate structure <slug> and sds validate brief against the written artifact.
  3. Stop once and ask for explicit approval.
  4. After approval, implement the complete brief without per-task approval stops.
  5. Run the brief's verification and exactly one compact independent review.
  6. If a mandatory full-lane trigger appears, stop and ask before escalating to the full lane.

Read the full file on GitHub · 289 lines

Changes

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.

  1. 3d ago First seen · 289 lines · 54 tokens per session scan A deb018d97bb8

Subscribe to this mod's changes

Spec-Driven is an agent published in the GitHub repository lindoelio/spec-driven-steroids (54 stars, last pushed 24d ago), licensed MIT. It adds 54 tokens to every session and 3,479 once invoked, about $0.0003 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 2026-08-30.