garden-health
Use when the user wants to maintain vault health, garden the knowledge base, check for orphans, find what is stale, audit structural gaps, or see a vault health report. Composes analysis tools into a full health picture before proposing any remediation.
- License 许可
- MIT
- 最近更新
- 2026-03-22
- Stars
- ★ 0
- Forks
- 0
01 前言 Front Matter
- name
- garden-health
- version
- 1.0.0
- description
- Use when the user wants to maintain vault health, garden the knowledge base, check for orphans, find what is stale, audit structural gaps, or see a vault health report. Composes analysis tools into a full health picture before proposing any remediation.
- disable-model-invocation
- true
02 资源文件 1
- 参考文档references/garden-audit.md2.7 KB
03 SKILL.md 正文
Garden Health Audit
Garden and maintain the vault by composing a full structural health picture before proposing any remediation. This skill uses the Fan-Out pattern: run all analysis tools in parallel, synthesize the report, then confirm with the user before any writes.
Iron Laws
Audit first, act second. Complete the full analysis (steps 1-5) before proposing any remediation. Never suggest fixes mid-audit — the full picture changes the priority order.
Never remediate without confirmation. Present the complete maintenance report and wait for explicit user approval before any writes. Orphan reweave, maturity promotion, and note creation are all write operations.
Check result.success after every MCP call. If any analysis tool fails,
note the failure, continue the remaining reads, and report the partial results.
Workflow
-
Read
ztlctl://garden/backlog— load stale seeds (age, last-modified) and orphan notes (link counts). This is the maintenance backlog candidates list. -
Read
ztlctl://review/dashboard— load the external review workbench snapshot: pending reviews, external sources awaiting processing. -
vault_review()— comprehensive aggregate snapshot: total notes, stale count, orphan count, maturity distribution, content type breakdown. Checkresult.success. -
graph_gaps(top=10)— find structurally isolated clusters with no inter-cluster edges. These are knowledge islands that may need bridging notes. -
graph_bridges(top=10)— find high-value bridge nodes whose removal would disconnect clusters. High-risk bridge nodes deserve protection or redundant linking. -
Synthesize health report — combine all reads into a prioritized summary:
- Orphan count and worst offenders (0 links, old age)
- Stale seed count and top candidates for promotion
- Structural gap clusters (isolated groups with no cross-links)
- Bridge nodes at risk (high bridge_score, low redundancy)
- External review backlog (pending sources from dashboard)
-
Present action list with confirmation gate — propose prioritized remediation with scope options:
- Connect orphans:
reweave(content_id="<id>")for each orphan - Promote stale seeds:
update_content(content_id="<id>", changes={"maturity": "budding"}) - Document gaps:
create_note(title="<bridging topic>")to link isolated clusters
Ask: "Process all / connect orphans only / promote seeds only / select specific items?" Wait for user response before any writes.
- Connect orphans:
Distinction from ztl:review-triage
ztl:review-triage focuses on the work queue: actionable items (tasks, stale notes, drafts needing updates). ztl:garden-health focuses on structural vault health: orphans, gaps, bridges, and maturity staleness across the full vault. When the user asks "what needs attention," clarify: work queue items or vault structure?
See references/garden-audit.md for tool output schemas and remediation option
details.
04 同仓库更多 Skills
vault-methodology
Use when working in a ztlctl vault — creating notes, organizing knowledge, tagging, linking, or making decisions about content types and structure. Activates for any knowledge management task.
synthesize
Use when the user wants to synthesize scattered notes into a unified artifact, consolidate knowledge on a topic, connect existing notes, find structural gaps, or create a synthesis note. Activates on "synthesize", "consolidate", "connect my notes on X", "summarize what I know about X", "find gaps in X", or "create synthesis".
session
Use when the user wants to start, begin, or open a work session, OR close, end, or finish an active session. Manages the full session lifecycle: pre-flight checks, polaris alignment, session start, and session close with enrichment pipeline.
session-workflow
Use when starting, running, or closing research sessions. Guides structured knowledge capture workflows: start → capture → synthesize → close.
review-triage
Use when the user asks to review the work queue, triage pending items, clear the backlog, or find out what needs attention in the vault. Surfaces prioritized actionable items and batch-processes approved updates.
review-contradictions
Use when the user wants to review conflicting notes, find inconsistencies in the vault, check for contradictions between knowledge claims, or inspect semantically-close notes that may conflict. Runs the per-pair evaluation loop to surface genuine contradictions and dismiss false positives.