feat: Phase D revised — memory rationalization, agent persistent memory, expanded verification (v2.8.0)#156
Merged
michael-wojcik merged 8 commits intomainfrom Feb 6, 2026
Merged
Conversation
Add memory: user frontmatter to 8 agent definitions, enabling cross-project persistent memory stored at ~/.claude/agent-memory/<name>/. The first 200 lines auto-load into each agent's system prompt, allowing specialists to accumulate domain expertise across sessions.
Rewrite "Future Executor: TeammateTool" section in pact-scope-contract.md as "Future Executor: Agent Teams" with real API tool names (TeamCreate, SendMessage, TeamDelete, Task with team_name/name). Add architectural notes on behavioral differences (no nested teams, peer-to-peer messaging, no conversation history inheritance). Update SSOT extract in pact-protocols.md and verify-protocol-extracts.sh line ranges.
Reduce MAX_WORKING_MEMORIES from 5 to 3 to limit token overlap with platform auto-memory. Add three-layer memory architecture documentation to CLAUDE.md, SKILL.md, and pact-memory-agent.md clarifying the complementary roles of auto-memory (session learnings), pact-memory (structured institutional knowledge), and agent persistent memory (domain expertise). Includes memory: user frontmatter for pact-memory-agent (combined with section 4 updates).
Add 26 new checks to verify-scope-integrity.sh across 7 gap areas: worktree integration (7), memory hooks baseline (3), executor interface (3), agent persistent memory (8), scope detection heuristics (3), completeness signals (1), Agent Teams docs (1). Total verification checks: 108 (was 82). All 4 scripts pass.
Add 4 integration test scenario runbooks and a 3-tier canary verification checklist to testing/ (repo-level development infrastructure, not part of the distributed plugin): Scenarios (P0/P1): - Standard orchestration (P→A→C→T lifecycle) - Memory layer interaction (auto-memory + pact-memory + agent memory) - Worktree lifecycle (setup → work → cleanup) - Scoped orchestration (detect → atomize → consolidate) Canary checklist: - Tier 1: Automated (run 4 verification scripts) - Tier 2: Structural (reviewer cross-reference checks) - Tier 3: Behavioral (dogfooding via PACT self-orchestration)
Reserve v3.0.0 for full Agent Teams integration. This release adds memory layer rationalization, agent persistent memory, expanded verification scripts, and integration test scenarios.
michael-wojcik
force-pushed
the
feature/phase-d-revised
branch
from
72f52d0 to
3af339e
Compare
michael-wojcik
changed the title
- Fix agent count "7" to "8" in memory layer test scenario - Fix verification script count "three" to "four" in standard orchestration scenario - Update stale "Last 5" to "Last 3" in session_init.py template - Hoist AGENTS_DIR variable to top of verify-scope-integrity.sh - Add unit test for MAX_WORKING_MEMORIES entry count limit
This was referenced Feb 6, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Phase D (Revised) of the PACT v3 plan — focused on memory layer rationalization, documentation updates, and regression safety expansion.
memory: userpersistent memory to all 8 PACT specialist agentsTest plan
bash scripts/verify-protocol-extracts.sh— all 16 checks passbash scripts/verify-task-hierarchy.sh— all 24 checks passbash scripts/verify-scope-integrity.sh— all 68 checks passmemory: userin all 8 agent frontmatters:grep -l "memory: user" pact-plugin/agents/*.md | wc -l= 8