agentsclimarketplace

Openplan garden

Skill BANANASJIM/openplan-skill/skills/openplan-garden

Portable OpenPlan-style Agent Skills for Codex and Claude Code: intent alignment, docs governance, review, handoff, and safe auto workflows.

Install
npx -y skills add BANANASJIM/openplan-skill --skill openplan-garden

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

  • 4 stars4 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.

What its author says it does

Copied from the file, not written here

Inspect and improve documentation quality using OpenPlan-style document gardening. Use when Codex or Claude Code must check memory, docs, design notes, handoff snapshots, review reports, specs, or generated documentation for goal/intent drift, broken structure, missing rationale, stale claims, duplicated content, unclear authority, poor future-agent readability, zero-context comprehension gaps, or periodic documentation maintenance. This skill is platform-neutral and read-only by default; local link checkers and linters are optional evidence sources, not required gates.

SKILL.md

4.0 KB, as published. Nobody here has run it

OpenPlan Garden

Use this after $openplan-core for document quality, periodic documentation maintenance, and semantic drift inspection. Garden is about preserving the clarity of human goal/intent over time.

Core Rule

Garden first reports. It does not rewrite durable docs or memory unless the user explicitly asks for a cleanup/edit pass.

What Garden Checks

  1. Goal/intent clarity: can a future human or agent tell why this exists?
  2. Authority clarity: is this source of truth, draft, handoff, review, or temporary brief?
  3. Layer clarity: does it mix goal/intent, design, spec, implementation, and review without labels?
  4. Rationale: are decisions explained, not just stated?
  5. Staleness: are time-sensitive or superseded claims marked?
  6. Duplication: is content copied where a reference would work?
  7. Traceability: are important claims tied to sources or evidence?
  8. Zero-context readability: can a competent reader understand it without the chat?
  9. Human-decision boundary: does it record human decisions separately from agent suggestions?
  10. Next-action clarity: does it say what should happen next, if anything?

Periodic Garden Pass

Use garden as the periodic documentation cleanup workflow. A periodic pass checks:

  • stale or superseded documents;
  • duplicated rationale that should become references;
  • missing or unclear authority labels;
  • memory/docs conflicts;
  • abandoned TODOs or unresolved questions;
  • research/review/handoff artifacts that should be archived, promoted, or left alone;
  • docs that no longer pass zero-context comprehension.

Garden first reports proposed cleanup. It may edit only when the user explicitly asks for a cleanup pass, and durable goal/intent changes still require $openplan-record and human confirmation through the human-facing coordinator.

Workflow

  1. Identify document type: memory, design, decision, spec, handoff, review, tutorial, temporary brief.
  2. Identify the docs root when a whole documentation surface is being reviewed.
  3. Load only the document and directly referenced context needed to judge it.
  4. Run optional deterministic checks if available and safe.
  5. Apply references/checklist.md.
  6. Report findings using references/report-contract.md.
  7. If the user requests cleanup, use $openplan-record before editing durable memory/docs.
  8. For periodic maintenance, separate safe mechanical cleanup from human-owned meaning changes.

Docs Root

Garden can operate on any docs root: .openplan/, docs/, repo root, a separate docs repo, or a path supplied by the user. Treat document-map paths as relative to the selected root. If the root is ambiguous, report the ambiguity before scanning broadly.

Severity

  • G0: misleading human goal/intent, wrong source of truth, or future agent likely to act incorrectly.
  • G1: missing rationale, stale unmarked claim, authority/layer confusion, or broken decision boundary.
  • G2: duplicated content, weak traceability, incomplete next action, or zero-context readability gap.
  • G3: wording, formatting, title, or navigation improvement.

Optional Checks

Use deterministic tools only when available:

  • markdown link check;
  • local docs linter;
  • grep for TODO/stale/duplicate markers;
  • template section checks.

If not available, continue with semantic review and state checks not run.

Claude Code

Use assets/claude-code-command.md as a slash command or prompt seed.

Keep looking

Skills are one crate of 328,083. 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.