agentsclimarketplace
Sign in

Grizzly init

Skill HarishDvs/Grizzly/skills/grizzly-init

Set up a new Grizzly writing project. Use when the user wants to start using Grizzly on a novel, when no NOVEL.md exists in the project, or when the user asks to onboard an existing manuscript. Creates NOVEL.md, VOICE.md, STANDARDS.md and the codex scaffold through a guided interview.From its SKILL.md

Install
npx -y skills add HarishDvs/Grizzly --skill grizzly-init

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

3.8 KB, 849 tokens by cl100k_base, as published. Nobody here has run it

Grizzly Init — Project Setup Wizard

Creates the project state files through a short interview plus analysis of the author's own writing. Do the steps in order. Don't skip the voice derivation; it is the highest-leverage step in the whole suite.

Step 1 — Locate the project

Ask where the manuscript lives (or will live) and confirm the filename pattern for chapters. All state files go at the project root. If state files already exist, stop and ask whether to update instead (route to grizzly-voice for VOICE.md).

Step 2 — The interview (NOVEL.md)

Ask, in one batch, only what the template needs:

  1. Title, one-to-two-sentence premise, current status.
  2. Genre and subgenre.
  3. Format: serial or traditional novel. Explain why it matters in one line (serials need per-chapter hooks and a shorter payoff cadence).
  4. POV scheme and tense.
  5. Benchmark titles: 2-4 published works the author wants their scenes measured against for amplitude. If they hesitate, ask what they read and admire in-genre.
  6. Anything the assistant must never do in this project (the never-do list).

Fill templates/NOVEL.md and write it to the project root. Show it; let the author correct before saving.

Step 3 — Voice derivation (VOICE.md)

Ask for 2-3 passages of the author's own prose, 300+ words each, that they consider their best representative work. Published or not doesn't matter; AI-assisted passages do (decline them politely; the spec must be derived from the author's hand).

Analyze the passages for every field in templates/VOICE.md:

  • Sentence-length distribution (count a sample; don't guess). Fragment frequency and function.
  • Paragraph shape and scene-transition style.
  • Narrator distance, and what the narrator never does in these samples.
  • Dialogue: information load, tag style, subtext habits.
  • How emotion enters (hands, objects, silence, action). Cite the actual instances.
  • How violence or high intensity enters, if present in the samples.
  • Register and diction: words that recur, words that would break the texture.
  • Imagery density and source domains.

Then propose 3-6 anchor lines from their samples and ask the author to confirm or swap them. Write VOICE.md. Tell the author this file is theirs to edit and can be re-derived later with grizzly-voice as their style evolves.

If the author has no samples yet (brand-new writer), say plainly that VOICE.md will start thin: seed it with their preferences by interview, mark it provisional, and re-derive after they have 2-3 chapters of their own.

Step 4 — Standards (STANDARDS.md)

Copy templates/STANDARDS.md with its seeded defaults. Walk the author through the nine defaults briefly; let them strike or amend any. Add their project-specific rules to Author Additions. Save.

Step 5 — Codex scaffold

Create codex/ with chapters/, arcs/, characters/, and an _Index.md from the codex templates. If a manuscript already exists, offer grizzly-cards bootstrap mode to card it chapter by chapter, and note the cost saving (cards make long manuscripts cheap to work on; bootstrap once, then maintain by the drift rule).

Step 6 — Genre pack

If a pack under packs/ matches the declared genre/format, record it in NOVEL.md. Packs adjust defaults (pacing norms, hook conventions); they never override VOICE.md.

Step 7 — Close

Summarize what was created, in one short list. Tell the author the entry points: draft, edit, review, audit, plan, brainstorm, cards. Do not demo anything on their manuscript unasked.

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.

Gives 0 of the 12 instructions most project setup skills give in 849 tokens

Counted across 1,553 of the 3,091 authors here whose files we hold, read 2026-09-06

  • Write the configuration filein 36 of 1553
  • Create the directory structurein 35 of 1553, across 33 files
  • Verify the setupin 31 of 1553, across 28 files
  • Run the setup scriptin 30 of 1553, across 29 files
  • Pre-determine the required sample sizein 29 of 1553, across 12 files
  • Check if the configuration already existsin 29 of 1553
  • Document every testin 26 of 1553, across 10 files
  • Start with a hypothesisin 26 of 1553, across 11 files
  • Ask one question at a timein 22 of 1553
  • Test a single variable per testin 21 of 1553, across 9 files
  • Read product marketing context before asking questionsin 19 of 1553, across 8 files
  • Do not peek and stop earlyin 18 of 1553, across 7 files

Said here and by no other author read

  • Ask where the manuscript lives
  • Fill templates NOVEL md and write it
  • Ask for passages of the author's prose
  • Copy templates STANDARDS md
  • Create codex scaffold

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.

Keep looking

Skills are one crate of 325,949. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.