Scan expired documents under .local.claude and archive superseded files.
npx skills add https://github.com/ejoongseok/claude-settings --skill garden-ejoongseok
Please help me install this Agent Skill. Skill: garden Source: https://github.com/ejoongseok/claude-settings/tree/main/claude-code/skills/garden Command: npx skills add https://github.com/ejoongseok/claude-settings --skill garden-ejoongseok
It prevents your growing local knowledge garden from becoming stale by automatically identifying expired documents, harvesting missed insights into durable sources, and archiving what’s no longer needed.
.local.claude/ (excluding archive/) using path-based categories and optional frontmatter retention..local.claude/biz-rules.md, modules/*.md, customers/*.md, team.md, people/*.md) and produces a report of what must be reflected..local.claude/archive/YYYY-MM/..., updates frontmatter, and refreshes .local.claude/INDEX.md (kept within a 60-line limit).Run the garden skill to generate a report for expiring documents under .local.claude/, review the harvest and archive plan, then approve execution to apply updates and archive the expired files.
High-intent search queries and answers about installing and using this skill.
Automate document lifecycle cleanup by scanning `.local.claude` paths for expired items using mtime and line counts. It archives superseded files after generating an approval report, ensuring your Claude context management stays current without manual deletion.
Knowledge harvesting identifies missed insights in expired files by comparing them against durable sources like biz-rules and modules. It generates a report of reflections needed, updating durable documents before moving expired items into an archive directory.
Archive stale `.local.claude` files by running an expiration scan that prioritizes frontmatter retention when present. After approval, it moves files to `archive/YYYY-MM/` with `archived/harvested_to` metadata and safely refreshes `INDEX.md`.
Yes, Claude context management archiving requires a root `SKILL.md` YAML name and description. This configuration drives the category-based scanning logic for daily, meetings, learn, reports, cs, memo, and docs-parsed workflows.
Automated index updating has a strict 60-line limit for `INDEX.md` and requires user approval before applying any harvest or archive actions. It excludes files already within the `archive/` directory from expiration scanning.
pradeepmouliGenerate baseline-spec.md and current-state.md from repository analysis.
starlink-awakenOrganize and query the PAI Knowledge Archive across four entity domains.
timominkusConvert transcripts or notes into structured Obsidian meeting records and update hub/daily files.
aryaniyapsExtract and convert webpage content into markdown using firecrawl CLI.