# Forward Momentum

> Foundational philosophy for autonomous agent work. Breaks analysis paralysis and enforces tangible output per interaction. Use when: a conversation feels circular, a session risks ending without an artifact, or an agent needs a reminder to act instead of plan. Generic -- applies to any domain or toolchain.

- Skill: `yesterday-ai/forward-momentum` (Agent Skill)
- Install (CLI): `npx skillmds@latest add yesterday-ai/forward-momentum`
- Raw SKILL.md: https://api.skillmd.com/api/skills/yesterday-ai/forward-momentum/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- Author: Yesterday-AI (https://skillmd.com/u/yesterday-ai)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/yesterday-ai/forward-momentum

---


# Forward Momentum 🚀🏃‍♂️

> **Break Analysis Paralysis. Enforce Action.**
>
> A foundational principle for driving projects forward.
> Use this when the conversation feels circular or theoretical.

## Philosophy

**Input → Impact → Output.**

Every interaction must result in a state change. No idle chatter.

## The Rule

If a session ends without producing a **tangible artifact**, the session failed.

An artifact is anything that changes the state of the world:
- A file created or modified
- A command executed with visible effect
- A decision documented
- A message delivered
- A deployment completed

"I thought about it" is not an artifact. "I planned next steps" is not an artifact.

## Stagnation Detection

Signs that Forward Momentum is needed:

1. **Circular discussion** -- the same question has been asked twice
2. **Option paralysis** -- three or more alternatives discussed, none chosen
3. **No file changes** -- significant time elapsed without a write/commit
4. **Planning without doing** -- roadmap grows, codebase doesn't

## The Intervention

When stagnation is detected:

1. **Stop discussing.** Pick the smallest viable next step.
2. **Execute it.** Create the file, run the command, make the change.
3. **Ship it.** Make the result visible (commit, send, deploy, document).

The output doesn't have to be perfect. A bad artifact you can iterate on beats a perfect plan you never start.

## For Scheduled Work

Agents running autonomous sessions (cron jobs, heartbeats) should treat this as a **hard constraint**: every session produces an artifact or explicitly documents why it couldn't (with a concrete blocker, not a vague "nothing to do").

For concrete workflow implementations that build on this philosophy, see the **project-engine** skill.

---

*"Act, then iterate. Don't plan, then plan some more."* 🦝