Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion .claude-plugin/plugin.json
Original file line number Diff line number Diff line change
@@ -1,6 +1,6 @@
{
"name": "agent-mem",
"version": "0.6.0",
"version": "0.7.0",
"description": "Bounded project memory tools and seven deterministic capture hooks, without a resident worker.",
"mcpServers": {
"agent-mem": {
Expand Down
5 changes: 5 additions & 0 deletions CHANGELOG.md
Original file line number Diff line number Diff line change
Expand Up @@ -2,6 +2,11 @@

## Unreleased

## 0.7.0

- `amem setup <claude|codex|kimi|pi|all>` (agent-mem#40 F2, F3, AC2, tasks #417, #432): a dry run by default, and a person-only `--apply`/`--undo` that registers the MCP server, hooks and a SessionEnd ingest through each client's own registration where it exists, idempotent, with semantic undo that keeps the client's own later edits and never touches AGENTS.md, CLAUDE.md or rule files. The registered runtime is the PATH-resolved `bun` (not a version-pinned Cellar path), and a temporary package entry is refused. The Claude Code plugin manifest gains the inline MCP server and a marketplace entry, so `claude plugin install` works without `--plugin-dir`. Advisory plugin-registry checks never block preview, apply or undo.
- Quick start and an installed-package CI test (agent-mem#40 F5, AC1, task #420): the README opens with install, `amem setup`, the review loop and the ingest cadence. CI packs the tarball, installs it offline into a clean HOME on macOS and Linux, and runs the installed `amem setup all` dry run (the whole HOME is snapshotted and must not change), MCP initialize and tools/list, one search and one hook with a durable capture record. No output may name ahub.
- CI timing (agent-mem#48, task #434): the brief deadline test runs on an injected clock, and the hooks and MCP budget scripts retry one isolated stalled sample under unchanged limits, fail on a repeated stall, and count stalled attempts per group. RSS, protocol and correctness failures never retry.
- Recall validity and intent briefs (agent-mem#40 F10-F12): current public reads omit superseded/expired notes; person history/as-of reads retained state under current forget/prune masks. Added advisory path staleness, shared fail-first byte-capped brief/plan-guard exports and `amem brief`/`amem_brief`. Machine-local nullable usage observations record rendered output only after successful emission; delivery/citation remain explicit host callbacks. Historical promotion timing is reported unknown only for emitted promoted rows, never fabricated. All exported historical APIs require explicit person authority; default graph is live-only. Recall uses one guarded policy/source snapshot, with a reproducible latency probe in scripts/measure-recall-latency.ts. Hub adoption is tracked separately as agent-hub backlog card 436.

- Agent candidate tools and person-reviewed lifecycle (agent-mem#40 F8/F9): explicit and structured fail proposals, revision-bound `amem_correct`, goal-linked promotion candidates accepted with the existing canonical promotion primitive, and `amem decay` preview/confirmed demote or explicit forget. Current-profile review supports shared-v1, generation and split homes; agent authority and shared-v1 schema/hash remain unchanged. MCP correction targets only its bound project and rejects an oversized complete accepted body at proposal. Decay receipts show exact expiry and partial failures report completed/uncertain/remaining ids in human or JSON output. Decay forget refuses unreviewed legacy-copy expansion before any batch change; ordinary person forget is unchanged. Compatibility for the next version: amem 0.6.0 and older candidate listings throw invalid candidate kind after a promotion row is offered; upgrade all readers/writers of that machine-local candidates.db before using this feature.
Expand Down
2 changes: 1 addition & 1 deletion README.md
Original file line number Diff line number Diff line change
Expand Up @@ -2,7 +2,7 @@

Curated memory with append-only JSONL logs, a derived SQLite index, short- and long-term tiers, a typed graph, goals with leases, and git sync between machines. Extracted from [agent-hub](https://github.com/STAIxBWLB/agent-hub), with compatible storage and entry IDs.

Version: 0.6.0. Requires Bun 1.4.2 or later; no runtime dependencies.
Version: 0.7.0. Requires Bun 1.4.2 or later; no runtime dependencies.

## Quick start

Expand Down
2 changes: 1 addition & 1 deletion package.json
Original file line number Diff line number Diff line change
@@ -1,6 +1,6 @@
{
"name": "@staix/agent-mem",
"version": "0.6.0",
"version": "0.7.0",
"description": "Curated, shared memory for coding agents: append-only logs, a derived index, tiers and a typed graph.",
"license": "MIT",
"type": "module",
Expand Down
Loading