Convert AI coding sessions into evidence-based retrospectives with prioritized improvements.
npx skills add https://github.com/iliaal/whetstone --skill ia-reflect
Please help me install this Agent Skill. Skill: ia-reflect Source: https://github.com/iliaal/whetstone/tree/main/plugins/whetstone/skills/ia-reflect Command: npx skills add https://github.com/iliaal/whetstone --skill ia-reflect
ia-reflect prevents low-discipline agent work by systematically turning a completed session into concrete, prioritized lessons that improve future performance instead of letting context reset erase what was learned.
remember: markers.Ask your agent to run /ia-reflect to review what went well and what went wrong in the last session, then capture the improvements it recommends for future chats.
High-intent search queries and answers about installing and using this skill.
To run an AI coding session retrospective, ask your agent to extract mistakes, friction, wasted effort, and wins with cited evidence. It then enforces actionable prioritized improvements with measurable success criteria and saves approved items to project-scoped memory files.
Yes, you can audit inbound and outbound PR and MR comment patterns to detect review-trap patterns and harvest heuristics. This review-aware audit extracts high-leverage insights using a 5-minute filter and applies them as actionable process improvements.
The best way to verify agent skills is to run a skill-level audit that checks success criteria, flags token inefficiency, and identifies missing edge cases. It proposes diff-ready fixes and requires an apply decision before modifying the audited skill.
Yes, it captures direct remember markers and approved operational learnings into project-scoped memory files. By extracting high-leverage insights using a 5-minute filter, it ensures memory capture prevents low-discipline agent work from disappearing after context resets.
You need to generate new skills when pattern detection identifies repeated task clusters that no existing skill covers. This occurs during a session retrospective when the agent audits invoked skills and recommends new skills to fill operational coverage gaps.
YanCheng-goSummarize completed session work and update MEMORY.md with progress and blockers.
opsMachineStore facts, decisions, or instructions into project or global memory files.
TwoTonedCapture and organize project learnings across gstack sessions for retrieval.
ldk-hubAnalyze completed epic stories to extract patterns, root causes, and lessons learned.