Rohit Ghumare

Tracking: native-surface audit for the 5 agents added in PR #677 (skills/prompts/rules, not just MCP)

1

Tracking issue. PR #677 shipped MCP-only adapters for 5 new agents (Warp, Cline, Continue, Zed, Droid). MCP is the lowest-common-denominator wrapper. It flattens agentmemory's 8 skills + 4 commands +...

PR #677 ships only `~/.config/zed/settings.json` (`context_servers:`) for Zed. Zed has a much richer auto-load surface we should use. Per Zed's [rules docs](https://zed.dev/docs/ai/rules), Zed auto-loads (in priority order) at project root:...

## Problem `mem::context` (the function `SessionStart` / `PreCompact` / `PreToolUse` hooks call to build the injected context block) reads from a subset of the memory tiers: - ✅ Pinned memory...

feat(website): add hero copy prompt button

2

## Summary - Add a secondary copy prompt CTA beside get started on the marketing homepage hero - Copies an onboarding prompt referencing llms.txt and AGENTS.md, with preview text derived...

From discussion #633: users on a fresh install report graph nodes growing fine but every consolidation tab (lessons, crystals, semantic facts, procedural skills, insights) stays at 0. Investigation: the consolidation...

Add diagrams and animations to course website (inspired by makingsoftware.com)

6

I love the work of the creator of MakingSoftware, and I used his design style for the website. Building on that direction, we should keep adding striking visuals and animations....

fix(cli): install pinned iii to private dir, fall back on PATH mismatch

2

## Problem Reported in #752. Fresh global install of `@agentmemory/agentmemory` on a box that already has `iii-engine v0.16.1` on PATH refuses to boot because agentmemory hard-pins v0.11.2: ``` ■ iii-engine...

## Problem We don't measure how often agentmemory injects the right memory but the agent fails to use it. Without this metric, "improve retrieval" and "improve reader behavior" are indistinguishable...

feat(smart-search): index semantic facts, procedural skills, crystals, insights, facets

1

## Problem `mem::smart-search` indexes compressed observations + lessons. It does not index: - `KV.semantic` (semantic facts produced by consolidation pipeline) - `KV.procedural` (procedural skills extracted from repeated patterns) - `KV.crystals`...

## Problem agentmemory has zero retrieval-quality benchmarks. Every change to the consolidation pipeline, embeddings, reranker, smart-search ranking, or context-injection budget split is shipped on intuition. No baseline to compare against....