agentleFS
vibe-cokitvibe-cokit/claude-code/skills/vk-hoh-plan/SKILL.md

vk-hoh-plan

Skill1 starsChanged 4 months agoApache-2.0View on GitHub

⚡ HoH Project Planner — synthesize S + prior evidence into one bounded dev-doc

The file

We split this skill into the rules it gives an agent

4 rules. About 0 are common, 2 are distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Read-only on A_{t-1} (Read/Grep/Glob + git status only).

    line 31
  2. Distinctive:

    IMPORTANT: No silent scope reduction — if the spec demands X, deliver X or split it across loops.

    line 59
  3. Distinctive:

    IMPORTANT: Read-only on the artifact — if the code looks wrong, state it as a task for the Developer.

    line 62
  4. Very specific:

    Diagram hook (opt-in): if the Runtime signals a visual helps, produce an architecture diagram via the archify skill CLI (diagram-layer.md).

    line 66

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

HoH Plan — Project Planner

HoH Project Planner role. Read S (spec) + the previous evidence E_{t-1}, inspect the current artifact A_{t-1} (read-only), and write a development document D_t that defines ONE bounded, locally-complete increment for loop t. You do NOT modify the artifact. HoH constrains the output (D_t schema), not your reasoning process.

SSOT: skills/vk-hoh-run/references/runtime-contract.md (schema, permissions table, reuse principle). Discipline: workflows/development-rules.md (Voice, Evidence Gate). For plan structure: activate the planning skill; for domain needs (Godot/asset/UI/…) activate domain skills as S requires — progressive disclosure.

<input>$ARGUMENTS</input>

Inputs (resolve from <input>)

  • S = <run-dir>/spec.md. E_{t-1} = <run-dir>/loops/loop-(t-1)-evidence.md (omit if t=1 → E_0 = ∅). A_{t-1} = project workspace (read-only). loop = t (from --loop or max existing loop + 1). Missing <run-dir>/--loop → ask ONE question.

Permissions (file gate — see runtime-contract.md › Permissions)

Read-only on A_{t-1} (Read/Grep/Glob + git status only); write ONLY D_t = <run-dir>/loops/loop-NN-dev-doc.md. The Runtime rejects your output if git status shows any source-tree change.

What D_t must contain

Reconcile S with E_{t-1}: address outstanding problems AND deliver one small concrete new capability (balance repair with growth). Pick ONE bounded, locally-complete objective — smallest surface that yields one observable behavior, including every related change needed to make it functional+testable; unrelated refactoring stays out. Reuse, not recreate — task the Developer to draw on existing skills/tools/MCP/asset pools for routine capabilities before rebuilding. Record preservation constraints (verified behaviors from E_{t-1} that must not regress). Write observable validation requirements (Given/When/Then, black-box from the product boundary — not "function returns true" but "player moves via WASD, camera follows"). Read A_{t-1} as implementation context so the objective reflects current structure.

Output — D_t (schema, all REQUIRED)

<run-dir>/loops/loop-NN-dev-doc.md per runtime-contract.md › D_t: frontmatter (loop, status: pending, preservation_constraints: [...]) + sections in order: Objective (≤6-word headline + 1 sentence) · Tasks (each naming files/components) · Preservation · Validation Requirements (Given/When/Then each) · Out-of-scope.

Return the D_t file path to the Runtime. Do not start implementation.

Notes

  • IMPORTANT: No silent scope reduction — if the spec demands X, deliver X or split it across loops; never "v1/placeholder/static for now".
  • IMPORTANT: Boundedness is a schema field — bundling unrelated changes gets rejected.
  • IMPORTANT: Read-only on the artifact — if the code looks wrong, state it as a task for the Developer; do not edit it yourself.
  • Delegate planning structure to the planner agent when the increment needs phase decomposition; this role owns the HoH-specific D_t contract, not generic planning.
  • Diagram hook (opt-in): if the Runtime signals a visual helps, produce an architecture diagram via the archify skill CLI (diagram-layer.md); never fork.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.