AGENTS.md · diff

git:20260824.62868b8 to git:20260912.33fd93e

4 added, 2 removed. Audit A to A.

# AGENTS.md

Context for AI coding agents working **on this repository**. If you are looking for how to

*use* the skill, read `skills/seedance-prompt-forge/SKILL.md` instead — that is the product.

## What this is

`seedance-prompt-forge` is a single Agent Skill: a procedure for writing, auditing, and

repairing prompts for Seedance 2.5 video generation. The distributable skill lives under

`skills/seedance-prompt-forge/`; repository-level files support development and release.

- `skills/seedance-prompt-forge/SKILL.md` — routing and universal rules. Loaded whenever the

skill activates. Keep under 500 lines and roughly 5,000 tokens.

- `skills/seedance-prompt-forge/references/*.md` — task-specific templates. Loaded on demand,

one level deep only.

- `skills/seedance-prompt-forge/scripts/lint_prompt.py` — deterministic checks. No third-party

dependencies. Python 3.8+.

- `dist/` — generated. Never edit by hand.

## Commands

```bash

- python3 scripts/run_tests.py # fixtures only

+ python3 scripts/run_tests.py # spec, distribution, fixtures, regressions

python3 scripts/run_tests.py --guide GUIDE.md # plus regression against the official guide

python3 scripts/build_bundle.py # regenerate dist/

./scripts/install.sh --list # show per-agent install paths

```

There is no build step for the skill itself. It is Markdown.

## Rules specific to this repo

**Fidelity to the source guide is the product.** Every template and rule traces to the

official Dreamina Seedance 2.5 prompt guide. When you add or change guidance:

- If the guide states it, state it the same way. Do not compress a structured block template

into prose — block structure is what users copy.

- If the guide does not state it, either leave it out or mark it inline as an assumption. The

`@Image` numbering note in

`skills/seedance-prompt-forge/references/reference-materials.md` is the worked example of how

to phrase an unsourced inference.

- Never present an inference as documented platform behaviour.

**The linter must never reject the guide's own examples.** `run_tests.py --guide` enforces

- this. Any new check has to survive all 29 filled examples with zero errors before it ships.

+ this. Any new check must survive all filled examples in the available guide export with zero

+ errors before it ships. Use `--expected-examples N` to assert extraction coverage; HTML exports

+ are supported alongside fenced Markdown. Keep local source exports out of distribution artifacts.

Warnings are acceptable where they represent a defensible tightening; errors are not.

**Regenerate `dist/` after touching the skill's `SKILL.md` or `references/`.** The bundle is what

chat-only surfaces consume; a stale bundle ships wrong instructions to ChatGPT and Claude

Project users.

**Keep frontmatter portable.** `name`, `description`, `license`, `compatibility`, `metadata`

only. `name` must match the directory name. Agent-specific fields such as `allowed-tools` are

ignored by most agents and are not worth the incompatibility.

**Version numbers live in `metadata.version`**, not as a top-level frontmatter key — a

top-level `version` is not in the Agent Skills spec and some implementations reject it.

## Scope boundary

This repo describes prompt construction. It does not call any generation API, ship model

weights, or reproduce generated media. Keep it that way — it is what lets the skill run

unmodified inside every agent without credentials or network access.