Skip to content

great_cto

v2.19.0 Feature

This release adds 2 notable features for engineering teams evaluating rollout.

Published 11d AI Agents & Assistants
✓ No known CVEs patched
Read the diff → Tool health → What is this tool? →

✓ No known CVEs patched in this version

Topics

agentic-coding claude-code-plugin claude-code-skills claude-code-subagents code-review cto
+2 more
multi-agent sdlc

Summary

AI summary

Token economy Phase 1+2 adds artifact‑summarization and task‑aware memory filtering to reduce LLM token usage.

Changes in this release

Feature Low

Adds artifact summary generation script generate-summary.mjs

Adds artifact summary generation script generate-summary.mjs

Source: llm_adapter@2026-05-28

Confidence: high

 —
Feature Low

Adds post‑tool‑use hook to auto‑generate artifact summaries

Adds post‑tool‑use hook to auto‑generate artifact summaries

Source: llm_adapter@2026-05-28

Confidence: high

 —
Feature Low

Adds pre‑push Git hook that blocks pushes with stale summaries

Adds pre‑push Git hook that blocks pushes with stale summaries

Source: llm_adapter@2026-05-28

Confidence: high

 —
Feature Low

Adds artifact‑summary contract document artifact-summary-contract.md

Adds artifact‑summary contract document artifact-summary-contract.md

Source: llm_adapter@2026-05-28

Confidence: high

 —
Feature Low

Adds 13 tests for summary‑enforce hook in tests/hooks/summary-enforce.test.mjs

Adds 13 tests for summary‑enforce hook in tests/hooks/summary-enforce.test.mjs

Source: llm_adapter@2026-05-28

Confidence: high

 —
Feature Low

Adds task‑aware memory filter script memory-filter.mjs

Adds task‑aware memory filter script memory-filter.mjs

Source: llm_adapter@2026-05-28

Confidence: high

 —
Feature Low

Adds memory‑filter prompt contract document memory-filter-prompt.md

Adds memory‑filter prompt contract document memory-filter-prompt.md

Source: llm_adapter@2026-05-28

Confidence: high

 —
Feature Low

Adds 18 tests for memory‑filter functionality in tests/memory-filter.test.mjs

Adds 18 tests for memory‑filter functionality in tests/memory-filter.test.mjs

Source: llm_adapter@2026-05-28

Confidence: high

 —
Feature Low

Updates architect agent to use filtered memory injection instead of tail/awk

Updates architect agent to use filtered memory injection instead of tail/awk

Source: llm_adapter@2026-05-28

Confidence: low

 —
Feature Low

Updates senior‑dev agent to use filtered memory injection instead of tail/grep

Updates senior‑dev agent to use filtered memory injection instead of tail/grep

Source: llm_adapter@2026-05-28

Confidence: low

 —
Full changelog

v2.19.0 — 2026-05-23

Added — Token economy Phase 1+2: artifact summaries + memory filter

Two-phase token-economy initiative (docs/plans/PLAN-token-economy-2026-q2.md).
Combined expected savings: 35-55% LLM tokens per agent start.

Phase 1 — Artifact summaries (30-50% savings on artifact reads)

  • scripts/generate-summary.mjs — Generates .summary.md (≤ 250 tokens) for every
    ARCH-*.md / PLAN-*.md / ADR-*.md / PHASE-*.md / QA-*.md / SEC-*.md /
    TM-*.md / RELEASE-*.md / PERF-*.md. Uses Anthropic Haiku → OpenRouter (Kimi K2) →
    deterministic heuristic. Idempotent. CLI: --all, --check, --force.
  • scripts/hooks/summary-enforce.mjs — PostToolUse hook: auto-generates .summary.md
    whenever an artifact is written. Fire-and-forget. Opt-out: GREAT_CTO_DISABLE_SUMMARY=1.
  • scripts/hooks/pre-push.sh — Blocks pushes with stale summaries.
    Bypass: GREAT_CTO_SKIP_SUMMARY_CHECK=1.
  • agents/_shared/artifact-summary-contract.md — Producer/consumer contract.
  • 13 tests in tests/hooks/summary-enforce.test.mjs.

Phase 2 — Task-aware memory filter (≥25% savings on agent startup context)

  • scripts/memory-filter.mjs — Filters lessons.md / decisions.md to the top-k
    entries most relevant to the current task. Same provider chain as Phase 1.
    Cost: < $0.0001/call. Latency: ~50ms heuristic, ~200ms Haiku.
    CLI: node scripts/memory-filter.mjs "<task>" <file> [--k=5] [--heuristic] [--stats].
  • agents/_shared/memory-filter-prompt.md — Prompt contract + integration table.
  • agents/architect.md — Replaces tail -100 / awk|head -60 with filtered injection.
  • agents/senior-dev.md — Replaces tail -50 / grep|head -40 with filtered injection.
  • 18 tests in tests/memory-filter.test.mjs. Opt-out: GREAT_CTO_DISABLE_MEMORY_FILTER=1.
View diff on GitHub

Weekly OSS security release digest.

The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.

No spam, unsubscribe anytime.

Share this release

Share on X Share on Bluesky

Track great_cto

Get notified when new releases ship.

Sign up free

About great_cto

Engineering-management layer of 34 specialist AI agents covering the full SDLC (architect, PM, senior-dev, reviewer, QA, security, devops, L3-support + 18 archetype-specific reviewers) with auto-detected archetypes and compliance gates (PCI-DSS, HIPAA, FedRAMP, GDPR, EU AI Act). Runs in Claude Code, Cursor, Codex CLI, Aider, and Continue via AGENTS.md + MCP. MIT.

All releases →

Related context

Earlier breaking changes

  • v2.32.0 Removed AgentShield scanner and its CLI commands.

Beta — feedback welcome: [email protected]