mirror of
https://github.com/rohitg00/agentmemory.git
synced 2026-09-14 20:16:33 +08:00
Trip
68fddd418e
* feat: OpenCode plugin with 22 auto-capture hooks (closes #156) - 22 hook handlers across session lifecycle, messages, tool lifecycle, parts, files, permissions, tasks, commands, and config - Two-layer enrichment pipeline: /context + /enrich via system.transform - Two slash commands: /recall and /remember - Full Claude Code hook parity documented with gap analysis * fix(plugin): use prompt_submit hookType so sessions get firstPrompt mem::observe checks hookType === "prompt_submit" to extract raw.userPrompt and set session.firstPrompt. The plugin was using "user_prompt_submit" which didn't match, so sessions were never named. * fix: address CodeRabbit review feedback on OpenCode plugin - Use ctx.worktree for projectPath instead of opaque project.id - Add Array.isArray guards before output.system/.context.push() - Only delete stashed files after successful enrich POST - Defensive JSON.stringify for undefined tool inputs - Per-session Map-based dedup sets to prevent unbounded growth - Fix negative duration_ms when time.completed is unset - Validate props.file and enforce MAX_STASHED_FILES in file.edited - Deduplicate /summarize call on session idle - Buffer early config events until session.created flushes them - Move contextInjectedSessions.add after successful context fetch - Add DEBUG-gated error logging to network helpers - Guard config input.agent/mcp/provider against non-object types - Fix MCP badge count 44→51 in plugin README * fix: add language identifier to fenced code block in plugin README * fix: store OpenCode session title as summary/firstPrompt on creation * fix(plugin): add session instruction injection and consolidation pipeline (closes #233) Three gaps from the Claude Code plugin port sweep: - Inject agentmemory usage instructions (memory_save, memory_recall, etc.) into the system prompt on first turn via experimental.chat.system.transform, replacing the skills mechanism that OpenCode lacks - Call /crystals/auto and /consolidate-pipeline on session.deleted, mirroring Claude's CONSOLIDATION_ENABLED behavior - Document MEMORY.md vs AGENTS.md architecture comparison (two-hop file bridge vs one-hop direct injection) Gap A (SubagentStop) is unfixable — OpenCode's SubtaskPart type has no completion/result fields. Gap C (Claude MEMORY.md bridge) is intentionally skipped — OpenCode uses direct injection. * fix(plugin): address adversarial code review — 7 critical/high fixes - Guard instructions push with Array.isArray check + fix TOCTOU race by moving contextInjectedSessions.add() before the await - Session-scope stashedFiles via stashFor() helper (was process-global, could cross-contaminate concurrent sessions) - Fix tool prefix in instructions (agentmemory_memory_ not agentmemory_) - Check typeof string before pushing .context into system arrays - Change olderThanDays: 0 → 7 in consolidation fire-and-forget - Increase fire-and-forget timeout from 5s to 30s (consolidation takes minutes, 5s was guaranteed to abort) * fix(plugin): scope stashedFiles.delete(sid) not .clear() on session.deleted * fix(plugin): address round 2 adversarial review — 3 critical fixes - Fix session cross-contamination: file.edited and tool.execute.before now use props.sessionID/input.sessionID fallback instead of only activeSessionId, preventing subagent hijack of parent stash - Fix contextInjectedSessions regression from round 1: move add(sid) to after instructions push (synchronous) but before context fetch (async), so failed /context calls don't permanently skip injection - Fix Map memory leak: prune session entries (stashedFiles, seenSubtaskIds, seenToolCallIds) when session.status goes idle, preventing unbounded growth for crash-killed sessions * fix(plugin): address round 2 remaining medium issues - Enforce MAX_STASHED_FILES cap on chat.message (was missing, could grow unbounded from 50+ file-ref messages) - Remove activeSessionId fallback from session.deleted; log warning when both info.id and sessionID are missing instead of guessing - Add subtaskSetFor/toolCallSetFor lazy-init wrappers matching stashFor pattern; fixes dedup failures for subagents spawned without a preceding session.created event - Guard chat.params against missing input.model (TypeError crash) * fix(plugin): address round 3 adversarial review — 10 HIGH/MEDIUM fixes - Add safeSlice() helper replacing all unsafe (v as string || "").slice() calls; handles objects/BigInt/circular refs via try/catch JSON.stringify - Restore activeSessionId fallback on session.deleted (if both info.id and sessionID missing, fall back like every other handler) - Fix message.updated to check info.id before info.sessionID (matches session.created/session.updated pattern) - Add props.sessionID to message.part.updated fallback chain - Guard undefined callID and subtask ID (prevent dedup silently dropping all subsequent undefined-ID events) - Cap todo.updated at 100 entries (match session.diff's slice pattern) - Fix duration_ms || → ?? (0ms genuine duration no longer falsifies) - Fix extractErrorMessage || → ?? (0/false error values preserved) - Replace extractErrorMessage.slice with safeSlice in retry handler * fix(plugin): address round 4 adversarial review — 6 CRITICAL/HIGH fixes CRITICAL: - Revert message.updated session ID resolution: info.id is the message UUID, not session ID. In message.updated, info is the message object (has role/tokens/modelID), not session info like in session.created. Use props.sessionID || info.sessionID || activeSessionId instead. - Fix post_tool_use_failure -> post_tool_failure: server types.ts defines "post_tool_failure" (no "use_"); compress-synthetic.ts classifies by that exact string. All error observations were misclassified as "other" instead of "error". - Fix contextInjected ordering: move add(sid) after the context fetch completes, not before. If /context times out, session is no longer permanently marked injected, allowing retry on next transform. HIGH: - Add process.cwd() fallback for projectPath (was null, causing silent 400s on every REST call when no workspace) - Guard session.start against null activeSessionId with early return (was sending sessionId: null to API) - Fix duration_ms: use typeof number checks instead of || 0 defaults; missing timing data now correctly reports null instead of 0ms * fix(plugin): address final adversarial blockers --------- Co-authored-by: xuli500177 <62830942+xuli500177@users.noreply.github.com> Co-authored-by: Trip <5579540+cl0ckt0wer@users.noreply.github.com> Co-authored-by: Rohit Ghumare <ghumare64@gmail.com>
6.1 KiB
6.1 KiB
Roadmap
This is agentmemory's public 12-month roadmap. It covers Q2 2026 through Q1 2027. The roadmap is the source of truth for where the project is heading; anything significant that lands in main should trace back to an item here or a ratified issue.
Items shift as evidence changes. Each quarter we publish a short retrospective on what landed, what slipped, and why — attached to the release notes.
How to read this
- Shipped — landed in main and tagged in a release.
- Active — in-flight, has an open PR or issue owner.
- Planned — accepted scope for the quarter, not started.
- Candidate — under consideration, may defer.
Anything not on this list that a contributor wants to pursue is welcome — open an issue labeled roadmap and it gets triaged against the quarterly theme.
Themes
- Q2 2026 — Depth. Multimodal memory, more connectors, close out backlog from the v0.9 cycle.
- Q3 2026 — Breadth. Hook parity across more agents, community expansion, OpenSSF best-practices alignment.
- Q4 2026 — Trust. Enterprise features — SSO, audit export, RBAC, long-running deployment story.
- Q1 2027 — v1.0. Stability, LTS branch, semver freeze on the REST + MCP surface.
Q2 2026 — Depth (April – June)
Shipped so far in this quarter
- iii console docs in README with vendored screenshots (#157)
- Health severity gated on RSS floor (#158 / PR #160)
- Standalone MCP proxies to the running server (#159 / PR #161)
- Audit coverage for
mem::forget+ audit policy doc (#125 / PR #162) @agentmemory/fs-watcherfilesystem connector (#62 / PR #163)- Next.js website on Vercel (PR #164)
- CI publishes all three npm packages on release (PR #166)
Active
- Multimodal memory — content-addressed image store, vision-prompt compression, disk quota + refcount on eviction (#64, PR #111)
- Governance baseline — this file, plus
GOVERNANCE.md,CONTRIBUTING.md,MAINTAINERS.md,CODE_OF_CONDUCT.md,SECURITY.md
Planned
- GitHub connector (
@agentmemory/github-watcher) — sync issues, PRs, discussions as observations. Shares thePOST /agentmemory/observewire format with the filesystem connector. - OpenCode hook bus (#156) — wired 22 hooks covering all 12 Claude Code hook types: session lifecycle (create/idle/status/compacted/update/diff/delete/error), messages & prompts (chat.message, message.updated user+assistant, message.removed), tool capture (before + rich ToolPart lifecycle in message.part.updated), memory injection (context + enrich via system.transform), part tracking (subtask, step-finish, reasoning, file, patch, compaction, agent, retry), file enrichment pipeline (stash via tool.execute.before + file.edited + file parts), permissions (updated + replied), task tracking (todo.updated w/ priority), commands (command.executed), config & model tracking (config + chat.params). Plus 2 slash commands (recall/remember). See
plugin/opencode/. - Session replay UI in the real-time viewer — scrub the timeline, inspect per-observation payloads.
- Benchmark harness in CI — keep the 95.2% R@5 number honest across releases by re-running LongMemEval-S on every minor tag.
Q3 2026 — Breadth (July – September)
Planned
- Additional maintainer onboarding — at least one Maintainer from a different organization added via the process in
GOVERNANCE.md. This is a prerequisite for advancing past the foundation's Growth Stage. - Slack / Discord connector — third source in the connector family.
- OpenSSF Scorecard — enroll, reach a Silver-equivalent score. Badged in the README.
- Hermes integration hardening — reach parity with the OpenClaw plugin surface (session lifecycle + tool-use hooks).
- Knowledge graph query language — small DSL on top of
/agentmemory/graphfor multi-hop questions. - First conference talk — submit to KubeCon / LlamaCon / similar.
Candidate
- Cross-agent shared memory namespace. Currently each agent installs its own instance. This would let a Claude Code session and a Cursor session recall each other's observations via a shared mesh node.
Q4 2026 — Trust (October – December)
Planned
- SSO gateway — accept OIDC in front of the REST surface for team deployments.
- Audit log export — streamable tail to S3 / Loki / stdout for compliance pipelines.
- RBAC on memory scope —
project:read,project:write,governance:deleterole set. - Long-running deployment guide — first-class Docker, systemd unit, and launchd plist.
- Performance SLO — publish p50/p95 recall latency targets, enforce via the benchmark harness.
- Security audit — external review of the REST surface + mesh-sync path. Fund through LF if foundation acceptance lands before end of quarter.
Candidate
- Agent-to-agent memory handoff protocol — standardize what one agent can inherit from another's memory, complementing MCP.
Q1 2027 — v1.0 (January – March)
Planned
- REST + MCP surface freeze. Any break requires a major-version tag per
GOVERNANCE.md. - LTS branch
v1.x— 12-month security-fix commitment. - v1.0 release — full documentation pass, all roadmap items from prior quarters either shipped or formally deferred.
- Foundation membership — Growth → Impact stage application if adoption + maintainer diversity metrics justify.
Candidate
- Hosted reference instance for the community to benchmark against.
- Reference implementation in a second language (Rust or Go) for the MCP server — would expand the set of runtimes that can host agentmemory.
Out of scope
For transparency, these are deliberately not on the roadmap:
- A cloud-hosted agentmemory SaaS.
- Billing, subscription tiers, commercial licensing beyond Apache-2.0.
- Agent frameworks themselves — agentmemory is a dependency, not a replacement for the agent runtime.
Feedback
Anything on this list you disagree with, or think should move up / down — open an issue tagged roadmap. Quarterly themes are revisited with every quarterly retrospective.