Extract post-validate learnings and stage taxonomy-classified proposals for human review.
npx skills add https://github.com/kapilvirenahuja/garura --skill reap
Please help me install this Agent Skill. Skill: reap Source: https://github.com/kapilvirenahuja/garura/tree/main/core/components/plays/reap Command: npx skills add https://github.com/kapilvirenahuja/garura --skill reap
Reap turns completed engineering outcomes into structured, reviewable learning proposals, so teams can identify what should be added or updated in long-term knowledge without manually tracing raw evidence.
Run reap for a validated issue by invoking the command: /reap <issue>.
High-intent search queries and answers about installing and using this skill.
To extract post-validate learnings, the process reads validate verdict evidence and STM context baseline, delegating extraction to the knowledge-extractor agent in analyze mode to identify gaps for LTM enrichment.
Taxonomy-classified proposals are generated by reading the build trinity outcomes and producing a two-level learning taxonomy aligned to the knowledge base structure for human review in STM.
ADR drafts with impact blocks are generated as Tier 1 tier-aware outputs by reading completed build trinity outcomes and staging the tiered learnings into proposals.yaml.
Yes, post-epic learning extraction enforces a mandatory Tether/Vanish human checkpoint before evidence self-committing to ensure reviewed proposals are written under evidence/reap/.
Deterministic workflows for gap identification apply to enterprise intent-driven development flows where an epic has completed prepare/implement/validate and requires structured LTM enrichment.
When post-validate review yields zero proposals, the deterministic workflow supports tiered outcomes including zero-proposal runs, ensuring evidence records still reflect the completed build trinity state.
courtneyr-devSecure WordPress development by sanitizing input, validating data, and escaping output.
kapilvirenahujaNormalize extractor-specific proposals.yaml into reconciliation-proposals.yaml with taxonomy validation.
usecompaiReviews conversations to classify and save learnings into the shared brain via *_brain_learn_* tool.
ben-mad-jlpStore user preference for auto-allowing rough-draft proposals in session state.
