Promote approved taxonomy-tagged issue proposals into product LTM and ADR archives.
npx skills add https://github.com/kapilvirenahuja/garura --skill garura-enrich
Please help me install this Agent Skill. Skill: garura:enrich Source: https://github.com/kapilvirenahuja/garura/tree/main/core/components/plays/enrich Command: npx skills add https://github.com/kapilvirenahuja/garura --skill garura-enrich
It provides a safe, reviewer-controlled terminal step that converts tiered, taxonomy-tagged proposals from issue STM into outward-facing product LTM and optionally new ADRs, then archives the processed issue so it is not re-applied.
Use the garura:enrich skill to enrich learning proposals for a specific issue by running: /garura:enrich <issue>.
High-intent search queries and answers about installing and using this skill.
Promoting taxonomy-tagged proposals into product LTM requires a terminal enrichment stage that enforces tier-based reviewer checkpoints, applying only explicitly approved changes to the long-term memory artifacts.
Tiered approval for ADR promotion forces explicit per-entry approval for Tier 1 entries, while using batch-with-diff defaults for Tier 2 and Tier 3, ensuring only confirmed successful writes trigger STM archival.
Processing multiple directories in sweep mode runs the enrichment stage across all active issue directories simultaneously, automatically excluding archived and pending subtrees from the promotion pipeline.
Issue STM archival ensures idempotency by moving the entire processed issue into an archived subtree after confirmed successful writes, preventing the same proposals from being re-applied downstream.
You can use native proposal sources from distill, reap, codify, or decode evidence directories because the enrichment process normalizes them into a reconciliation format suitable for review without modifying the original source proposals.
If taxonomy completeness validation fails during LTM enrichment, strict write-boundary roots and approval completeness gating block the process, preventing unverified proposals from being promoted to product LTM artifacts.
kapilvirenahujaNormalize extractor-specific proposals.yaml into reconciliation-proposals.yaml with taxonomy validation.
VitaDynamicsCreate sequentially numbered Architecture Decision Records in docs/adr/.
kapilvirenahujaExtract post-validate learnings and stage taxonomy-classified proposals for human review.
kapilvirenahujaAnalyze merged PR diffs and issue STM evidence to generate up to two staged proposals.yaml entries.
