research-openspec

A research report about three different tools called OpenSpec, including a command-line tool that manages written software proposals, specifications, designs, and task lists for AI coding assistants.

In plain words
What is it for?
Use it as background when evaluating OpenSpec, comparing its related projects, or planning specification-led software development.
Why use it?
It clarifies which OpenSpec people mean and how the tools differ, reducing confusion when choosing or discussing them.

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/hundia/autospec/research-openspec
Clone the repo
git clone --depth 1 https://github.com/Hundia/autospec
Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 7,027 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 $0.00000 $0.07027
Opus 5 $0.00000 $0.03513
Sonnet 5 $0.00000 $0.01405
Haiku 4.5 $0.00000 $0.00703

Measured 2d ago against content hash 8f60bb626a14, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

research-openspec 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 2d 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.

agents/research-openspec.md · 586 lines

How it starts

The opening of the file, as written. The whole thing — 586 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Research Report: OpenSpec

Researcher: B Date: 2026-03-20


What Is OpenSpec

"OpenSpec" is not a single product. Three distinct things share or are associated with the name:

1. @fission-ai/openspec (npm package / Fission-AI project)

The primary, actively-maintained tool. Published by Fission-AI at github.com/Fission-AI/OpenSpec. Latest version: 1.2.0 (February 2026). A globally-installed CLI tool for spec-driven development (SDD). It manages structured markdown artifacts (proposals, specs, design docs, task lists) and generates tool-specific slash-command adapters for 24+ AI coding assistants including Claude Code, Cursor, Windsurf, Gemini CLI, GitHub Copilot, Continue, Amazon Q, Cline, RooCode, Codex, and more.

Critical insight: OpenSpec does NOT call any LLM itself. It is a pure file-management and instruction-generation layer. It produces local markdown files and injects context into the user's existing AI assistant via that assistant's native slash-command or skill system.

2. openspec-for-copilot (VS Code extension, by atman-33)

A community-built VS Code extension at github.com/atman-33/openspec-for-copilot. It brings spec-driven workflows into VS Code by sending compiled prompts to GitHub Copilot Chat (via sendPromptToChat() utility) or optionally to Codex Chat. Not an official Fission-AI product.

3. GitHub Copilot SDK (@github/copilot-sdk)

Not "OpenSpec" but directly relevant to the no-API-key model this report targets. Official GitHub SDK for embedding Copilot agent capabilities via JSON-RPC into any Node/Python/Go/.NET application.


LLM Integration Architecture

Path A: @fission-ai/openspec — The Prompt-File Delegation Model

OpenSpec uses zero direct LLM calls. Its architecture is:

openspec init
    ↓
Scans for existing AI tool directories (.claude/, .cursor/, .gemini/, etc.)
    ↓
CommandAdapterRegistry generates tool-specific files:
  Claude Code  → .claude/commands/opsx/<workflow>.md
  Cursor       → .cursor/commands/opsx-<workflow>.md
  Gemini CLI   → .gemini/commands/opsx/<workflow>.toml
  Windsurf     → .windsurf/workflows/opsx-<workflow>.md
  GitHub Copilot → .github/prompts/opsx-<workflow>.prompt.md
    ↓
Skills layer: .claude/skills/openspec-propose/SKILL.md (universal YAML-fronted markdown)

Read the full file on GitHub · 586 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. 2d ago First seen · 586 lines · 0 tokens per session scan A 8f60bb626a14

Subscribe to this mod's changes

research-openspec is an agent published in the GitHub repository Hundia/autospec (4 stars, last pushed 2mo ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 7,027 tokens. 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.