akitaonrails/ai-memory
67.8
Adequate · 14 November 2026
498.9k
lines of production code
Rust
primary language
5
measurements over time
What this system is
Features
Add Antigravity CLI (agy) hook support
Added a new set of shell and PowerShell hooks for the Antigravity CLI (agy) harness. These scripts integrate the CLI with the AI memory system by forwarding lifecycle events (session-start, pre/post-tool-use, pre-invocation, stop) to the memory server and handling session handoffs, ensuring that briefing data is correctly passed during session initialization.
hooks/antigravity-cli · high confidence
Add Codex and OpenCode hook bundles for AI memory integration
New shell and PowerShell hook scripts have been added for the Codex and OpenCode agents to integrate with the local AI memory service (defaulting to port 49374). These hooks automatically forward session events—such as tool usage, prompts, and session start/end—to the memory server. Additionally, the session-start hooks now support cross-agent handoffs, allowing context to be automatically retrieved and prepended to new sessions without human intervention.
hooks/codex, hooks/opencode · high confidence
Add Devin CLI integration hooks
Added shell and PowerShell hook scripts for the Devin agent to integrate with the AI memory system. These hooks capture key lifecycle events—including session start, end, tool usage, and user prompts—and forward them to the local memory server, enabling context persistence and handoff capabilities for Devin sessions.
hooks/devin · high confidence
Add Fedora RPM packaging for ai-memory
A new RPM spec file has been added to define the packaging structure for the ai-memory project on Fedora-based systems. This spec configures the build and installation process for the ai-memory binary, documentation, configuration files, and associated systemd service units, enabling users to install the software via the RPM package manager.
packaging/rpm · high confidence
Add GitHub Copilot CLI agent hooks for session and tool events
This change introduces a set of shell and PowerShell hook scripts in the \hooks/copilot-cli\ directory to integrate GitHub Copilot CLI with the AI memory system. The hooks cover key lifecycle events including session start, session end, tool use (pre and post), subagent start/stop, user prompt submission, and stop. For most events, the hooks forward event data to the local AI memory server. Notably, the session-start hook also synchronously fetches any pending cross-agent handoff data and injects it into the Copilot session context, enabling seamless context transfer between agents.
hooks/copilot-cli · high confidence
Add Poolside Agent CLI hook capture and finalize-session
Added shell and PowerShell hook scripts for the Poolside Agent CLI (Pool) to capture agent lifecycle events. These hooks forward JSON payloads for session-start, user-prompt-submit, pre-tool-use, post-tool-use, and stop events to the AI memory server, enabling context tracking and session finalization for this specific agent integration.
hooks/pool · high confidence
Add environment configuration template for ai-memory service
A new environment file (ai-memory.env) has been added to the packaging directory to support the ai-memory systemd service. This file serves as a template for configuring essential runtime settings, including authentication tokens, allowed hosts, LLM provider selection (e.g., Anthropic, OpenAI), API keys, reasoning effort levels, and custom operator headers.
packaging/env · high confidence
Add native Arch Linux (AUR) packaging
Users on Arch Linux can now install ai-memory via the Arch User Repository. Two packages are provided: \ai-memory\ builds the application from source using the local Rust toolchain, while \ai-memory-bin\ installs prebuilt binaries for x86\_64 and aarch64 architectures. The packaging includes systemd service files for both system and user-level execution, along with helper scripts to guide post-installation configuration.
packaging/aur · high confidence
Add system user and directory setup for ai-memory service
The packaging now includes configuration files to automatically create the 'ai-memory' system user and the /var/lib/ai-memory directory during installation. This ensures the service has the necessary permissions and file structure to operate correctly on Arch-based systems.
packaging/sysusers, packaging/tmpfiles · high confidence
Add systemd service units for ai-memory
Users can now deploy the ai-memory MCP server as a managed system or user service. The new ai-memory.service unit runs the server as a dedicated 'ai-memory' user with strict security hardening (PrivateTmp, NoNewPrivileges, ProtectSystem=strict) and manages state in /var/lib/ai-memory. The ai-memory-user.service unit allows running the server under the current user's context, reading configuration from standard XDG directories.
packaging/systemd · high confidence
Added macOS LaunchAgent for ai-memory
A new launchd user agent plist has been added to the packaging directory, providing the macOS equivalent of the systemd user service. This configuration ensures the ai-memory binary runs as an interactive, persistent background service that starts at login and keeps the process alive, with output directed to user-specific log files.
packaging/launchd · high confidence
Claude Code and Codex marketplace plugins now support ai-memory session capture
New shell hook scripts for the Claude Code and Codex marketplace plugins enable automatic session capture via the ai-memory binary. The hooks detect the running agent (Claude Code, Codex, Muse, or Droid) and forward lifecycle events to the ai-memory service, handling configuration for server URLs and data directories. The implementation includes robust fallback logic: it ignores relative paths for the ai-memory binary, omits empty server URLs to let the binary use its own config, and retries calls against a fallback local server if an older ai-memory binary (pre-3.0) rejects the request due to missing server-url flags.
plugins/claude-code/scripts, plugins/codex/scripts · high confidence
Claude Code hooks now support cross-agent handoffs and subagent capture
The Claude Code integration now includes hooks for session-start, session-end, tool-use, and subagent lifecycle events, enabling automatic context transfer between agents and the ability to drop nested subagent captures. The session-start hook specifically fetches and injects pending cross-agent handoffs into the session context, while new subagent-start and subagent-stop hooks allow the server to manage subagent session lifecycles for capture control. All hooks are implemented for both POSIX shell and PowerShell, default to port 49374, and ensure clean debug logs by emitting JSON acknowledgments.
hooks/claude-code · high confidence
First-party DeepSeek Harness (DSH) lifecycle hooks
Added shell and PowerShell hook scripts in the \hooks/dsh\ directory to capture the full lifecycle of the DeepSeek Harness integration. These scripts forward events—including session start, user prompt submission, tool use (pre/post), subagent start/stop, and stop—to the local AI memory server, enabling automatic context capture and workspace strategy resolution via \.ai-memory.toml\ markers.
hooks/dsh · high confidence
Grok Build CLI integration hooks added
New shell and PowerShell hooks have been added to the Grok integration to support the Grok Build CLI. These scripts intercept session lifecycle events (start, end, stop, compact) and tool usage (pre/post-tool-use, user-prompt-submit) to forward payloads to the local AI memory service. The implementation specifically handles Grok's output constraints by preventing handoff acceptance during session start and user prompt submission, while the post-tool-use hook retrieves and injects context into the model's view. It also includes logic to identify and ignore subagent payloads to prevent them from incorrectly taking the parent session's handoff.
hooks/grok · high confidence
Initial database schema and migration framework for ai-memory
The ai-memory store now uses a versioned SQL migration system (V01–V27) to manage its SQLite data model. The initial schema (V01) establishes core tables for workspaces, projects, pages, sessions, observations, links, and audit logs, with full-text search on pages and observations. Subsequent migrations introduce agent handoffs (V02), page decay/retention tracking (V03), vector embeddings (V04), and performance indexes (V05–V08). The schema supports multi-agent environments by enumerating supported agent kinds (V09, V11, V20, V25, V26) and enforces data integrity through workspace-project pairing triggers (V18). It also adds multi-user attribution (V14–V16), path-based search indexing (V17), and data-repair migrations for orphaned observations (V19, V27). Finally, it introduces a comprehensive auto-improvement subsystem with pending proposals, scheduler state, patch support, and rejection buffers (V21–V24).
crates/ai-memory-store/migrations · high confidence
Introduce NixOS module for ai-memory service
Users can now deploy the ai-memory MCP server as a managed systemd service on NixOS. The new \nix/nixos-module.nix\ provides declarative configuration options for enabling the service, binding host/port, and toggling the web UI (\enableWeb\) and read-only API (\enableApi\). It supports declarative TOML settings via \services.ai-memory.settings\ (with typed options for common keys like \allowed\_hosts\, \log\_level\, and \auth\) while securely handling secrets through \ageSecret\, \sopsSecret\, or \environmentFile\. The module also includes hardened systemd sandboxing via \nix/systemd-sandbox.nix\ to restrict capabilities and system access.
nix · high confidence
Introduce OIDC bearer token resolution and local root token provisioning
The CLI now supports authentication via stored OIDC device-flow tokens in addition to static tokens. A new \auth\_bearer\ module resolves the bearer for server requests by first checking for a static token, then loading and refreshing an OIDC token from \auth.json\ if needed. A new \local\_auth\ module provisions a root bearer token for the local install, storing it in \\<data\_dir\>/auth-tokens/default\ with restrictive permissions, and seeds hook credential files so lifecycle hooks can authenticate without passing tokens on the command line. The \http\_client\ module uses this unified resolution logic to ensure consistent authentication across all thin-client CLI subcommands.
crates/ai-memory-cli/src · high confidence
Introduce ai-memory-llm crate with multi-provider LLM and embedding support
The new \ai-memory-llm\ crate provides a unified interface for interacting with large language models and generating embeddings. It supports multiple providers including Anthropic (via API keys and OAuth subscriptions), OpenAI, OpenAI-compatible endpoints, Google Gemini, GitHub Copilot, Codex, and the Cursor Agent CLI. The crate handles complex authentication flows, such as token exchange for Copilot and delegated refresh for Codex, while enforcing strict security boundaries like allowlisted environments for the Cursor child process. It also includes a local embedding capability using an in-process BERT model, with capacity management to ensure interactive queries are not blocked by background document indexing.
crates/ai-memory-llm/src · high confidence
Introduce containerized CLI wrapper and homelab deployment tooling
Adds a new \bin/ai-memory\ bash wrapper and \bin/ai-memory.ps1\ PowerShell wrapper that invoke the \ai-memory\ binary inside a Docker container, handling volume mounts for configuration and project data, forwarding environment variables, and managing version checks and upgrades. Includes a \bin/deploy\ script for building and pushing the image to a homelab Docker host, with safeguards to prevent overwriting multi-architecture release tags with single-architecture builds.
bin · high confidence
Introduce macOS menu bar companion for server management
A new macOS menu bar application has been added to bundle and govern the AI memory server. This companion provides a menu extra with a status indicator and quick actions to install, start, stop, and restart the server via a launchd agent. It surfaces server health details (version, page counts, LLM/embedding status) and, crucially, displays fatal startup errors directly in the menu rather than hiding them in log files. The app includes a Settings window for configuring the server URL, data directory, and bearer token (stored in the macOS Keychain), as well as a dedicated Status window to view and copy detailed server output. Also in this area: 3 commits (3 fixes) fixing companions.
companions/ai-memory-macos/Sources/AIMemoryMenu · medium confidence · unverified
Introduce native API credentials and auto-improvement proposal management
The store layer now supports native \aim\_\ API credentials (with generation, rotation, and revocation) and a full auto-improvement proposal lifecycle (staging, approval, rejection, and conflict handling), alongside read-time belief-strength confidence and per-tier retention decay parameters.
crates/ai-memory-store/src · high confidence
Introduce read-only web browser for the wiki
This change adds a new read-only HTTP browser for the wiki, mounted under \/web\ on the same server that hosts the MCP endpoint. It provides a set of routes for browsing projects, pages, and their version history (including unified diffs), searching across projects with a hybrid ranker, viewing link graphs, and checking server health. The browser includes built-in login and change-password pages for human sessions, handles authentication redirects for HTML navigation, and supports base-path mounting with proper \\<base href\>\ injection for relative URL resolution. The implementation is deliberately read-only, with no editing or write APIs exposed through this surface.
crates/ai-memory-web/src · high confidence
Introduces a server-rendered web UI for the ai-memory wiki
Adds a new read-only web interface for browsing the knowledge base, featuring a base layout with dark mode support, scoped project search, and a link graph visualization. Users can now view page history with diffs, navigate namespaces, and inspect server health metrics. The UI also includes built-in authentication pages for signing in and changing passwords, as well as a root-only pending writes triage page for approving or rejecting auto-improvement proposals.
crates/ai-memory-web/templates · high confidence
Kimi Code agent hooks for memory continuity and handoffs
Added shell and PowerShell hook scripts for the Kimi Code agent to integrate with the AI memory system. These hooks forward lifecycle events (session start/end, tool use, stop, subagent start/stop) to the memory server and implement cross-agent handoff delivery. Specifically, the user-prompt-submit hook fetches pending handoffs and delivers a one-time project briefing on the first prompt of a session, ensuring context continuity between agents.
hooks/kimi-code · high confidence
LongMemEval retrieval benchmark harness with A/B testing and QA accuracy
The \evals/src/retrieval\ module introduces a complete end-to-end retrieval benchmarking harness for the LongMemEval dataset. It replays question haystacks through the real server ingestion pipeline (or a local grep baseline) and scores retrieval accuracy (hit@k, recall@k) against evidence labels. The harness supports A/B testing between a baseline and a candidate configuration, reporting deltas in accuracy, latency, and context token usage. It also includes an opt-in LLM-as-judge mode to measure end-to-end answer quality by synthesizing answers from retrieved snippets and grading them against gold answers, as well as instruments to decompose retrieval gaps via reachability analysis and ingestion mode ablations. Also in this area: 1 commit (1 fix) fixing evals.
evals/src/retrieval · medium confidence · unverified
New CLI commands for API keys, contamination audit, and auto-improve telemetry
The CLI now includes \ai-memory api-key\ to manage native \aim\_\ API credentials (add, list, rotate, revoke) via the \/admin/api-credentials\ endpoints, \ai-memory audit-contamination\ to report cross-project session mis-bucketing from the \/admin/audit-contamination\ endpoint, and \ai-memory auto-improve-report\ to display read-only telemetry for auto-improvement outcomes from the \/admin/auto-improve/report\ endpoint.
crates/ai-memory-cli/src/commands · high confidence
New Docker deployment assets and hardened container defaults
The docker directory now includes a complete set of deployment files: a multi-stage Dockerfile (pinning the builder to rust:1.95-slim-bookworm), a local development compose file, a production compose example, and templates for TLS termination via Caddy or Cloudflare Tunnel. The container image now runs as a non-root user, exposes port 49374 by default, and enables the web UI in the CMD. Security is improved by defaulting \AI\_MEMORY\_ALLOWED\_HOSTS\ to include \host.docker.internal\ (fixing macOS wrapper connectivity) and \AI\_MEMORY\_IN\_CONTAINER=1\ (suppressing unauthenticated-bind warnings when binding to 0.0.0.0). A new \.env.production.example\ documents all configuration options, including the new \EMBEDDING\_API\_KEY\ for separate embedding credentials and bearer-token auth for LAN deployments.
docker · high confidence
New ai-memory-wikisync companion for two-way team-wiki synchronization
A new standalone Rust companion tool, ai-memory-wikisync, is introduced to keep allowlisted wiki page families in sync between a running ai-memory server and a local repository directory. It supports one-way export, two-way sync (importing repository edits back to the server via MCP tools), and a read-only CI check mode. The tool enforces strict safety measures, including path validation to prevent directory traversal, conditional writes to prevent overwriting concurrent changes, and opt-in delete propagation. It also includes a git post-merge hook installer to automatically sync after merges. Also in this area: 4 commits adding/updating tests in companions.
companions/ai-memory-wikisync, companions/ai-memory-wikisync/tests · high confidence
New desktop installer companion for non-terminal users
A new Tauri v2 desktop application (ai-memory-setup) is introduced to provide a graphical installation and setup experience for ai-memory, targeting users who prefer not to use a terminal. The app detects existing ai-memory installations (checking standard paths, package manager directories, and the user PATH) and uses them if found; otherwise, it installs a bundled binary to the standard user location. It launches the \ai-memory setup --web\ process, displays the setup page in its own window, and ensures the setup process stops when the window is closed. The implementation includes specific fixes for Windows PATH handling to preserve environment variable expansions and ensures that both stdout and stderr from the setup child process are read concurrently to prevent blocking.
companions/ai-memory-desktop · high confidence
New evaluation harness for LLM provider A/B testing and fidelity scoring
The \ai-memory-eval\ CLI now includes a live A/B harness that compares two LLM providers on the production consolidation prompt. It runs fixtures through both a baseline and a candidate provider, scoring literal fidelity by checking which facts were kept or dropped by the observation projection and the model itself. A zero-LLM projection subcommand isolates the projection stage to report which facts reach the prompt, and a seeded long-session generator creates reproducible test cases with planted facts to measure recall and hallucination rates.
evals/src · high confidence
New macOS menu bar companion app for AI Memory
A new macOS menu bar application (AI Memory) has been added to provide a graphical interface for managing the AI Memory server. This companion bundles the \ai-memory\ runtime, manages the background LaunchAgent service, and offers menu options to open the Web UI, view status, and access configuration and log files. It is designed as a wrapper around the existing CLI tools, ensuring that user data (wiki, SQLite, config) remains in the standard user data directory and is preserved across app updates. Also in this area: 3 commits (3 fixes) fixing companions.
companions/ai-memory-macos · medium confidence · unverified
New plugin templates for Claude Code and Codex
Added template files for marketplace plugins that connect Claude Code and Codex to an ai-memory server. The new \ai-memory-hook.sh\ launcher script detects the agent (Claude Code, Codex, or Muse Code), validates the \ai\_memory\_bin\ option to prevent relative paths, and handles compatibility with ai-memory binaries older than version 3.0 by retrying with a fallback server URL. The templates also include READMEs for both agents detailing installation, capture behavior, and security boundaries, as well as a \status\ skill template that guides users to run \ai-memory doctor\ for diagnostics.
crates/ai-memory-cli/templates/plugin · high confidence
New read-only web UI with per-project access control and graph views
The web interface now enforces per-project read access on every surface, ensuring users only see projects they are authorized to view. It introduces a read-only wiki browser with new features including a link graph visualization (both global and per-project), page version history with diffs, and a root-only health panel for server diagnostics. The UI also supports session listings, pending triage for auto-improvements, and builtin login/change-password pages, all secured by owner-guarded sessions and private file modes.
crates/ai-memory-web/src/routes · high confidence
New web-based setup wizard and expanded configuration defaults
The CLI now includes a local web interface (setup\_web.html) that allows users to review an installation plan, select agents, and apply configuration changes via a browser before running the setup command. The default configuration has been updated to change the HTTP bind port from 7777 to 49374 and introduces several new settings: a host-header allowlist for DNS-rebinding defense, opt-in hooks for capturing assistant stop events and serving Gemini-safe MCP schemas, a server-side rate limiter for hook ingestion, and configurable retention policies for raw observations and hook ledgers.
crates/ai-memory-cli/templates · high confidence
PowerShell hook library adds robust path resolution and marker walking
The \hooks/lib\ directory now includes \ai-memory-hook.ps1\, a PowerShell library that implements core logic for the hook system on Windows. This script provides functions to resolve the current working directory from various agent payloads (including specific handling for Devin and Cursor), normalize user home paths to avoid trailing separator collisions, and walk directory trees to locate and parse \.ai-memory.toml\ configuration markers. It also includes logic to detect server profile routing and parse TOML keys, ensuring parity with the native Rust hook implementation for path and configuration handling.
hooks/lib · high confidence
Behavioural changes
15 commits (6 fixes) modifying companions/ai-memory-importer/src/agentmemory
A change to existing behaviour in companions/ai-memory-importer/src/agentmemory — 15 commits (6 fixs), 6 files: “fix(importer): skip fenced code when deriving a page title”; “test: gate the Unix-only service tests; fix macOS dead code and importer fmt”; “test(importer): rerun long vault pages against the real read limits”; and 12 more.
companions/ai-memory-importer/src/agentmemory · medium confidence · unverified
2 commits (0 fixes) modifying crates/ai-memory-test-support
A change to existing behaviour in crates/ai-memory-test-support — 2 commits, 1 file: “perf(dev): a self-contained build and a two-tier test loop on every platform”; “test: stop parallel tests from reusing "dead" ports (\#2030)” (refs \#2030).
crates/ai-memory-test-support · medium confidence · unverified
2 commits (1 fix) modifying hooks/kiro-cli
A change to existing behaviour in hooks/kiro-cli — 2 commits (1 fix), 10 files: “feat(agents): add version-aware Kiro CLI lifecycle hooks”; “fix(hooks): never repeat inbox-notice query keys; compaction starts claim nothing”.
hooks/kiro-cli · medium confidence · unverified
2536 commits (1062 fixes) modifying (repo-wide)
A change to existing behaviour in (repo-wide) — 2536 commits (1062 fixs), 21 files: “release: v1.17.0”; “fix(wrapper): preserve stdin across every output mode”; “docs(changelog): record the 3.3.0 config key documentation (\#1861)”; and 2517 more (refs \#1861, \#792, \#1391, \#701, \#496, …).
(repo-wide) · low confidence · unverified
3 commits (0 fixes) modifying companions/ai-memory-importer/src/claude\_mem
A change to existing behaviour in companions/ai-memory-importer/src/claude\_mem — 3 commits, 1 file: “test(importer): replace the 3.4.0 claude-mem layout on a real server”; “feat(importer): migrate claude-mem databases”; “test(importer): read claude-mem lists back from a real server”.
_companions/ai-memory-importer/src/claude\mem · medium confidence · unverified
36 commits (16 fixes) modifying companions/ai-memory-ts
A change to existing behaviour in companions/ai-memory-ts — 36 commits (16 fixs), 52 files: “feat(mcp): cap memory\_read\_page bodies and shrink session-observation defaults”; “docs(changelog): reference PR 1816”; “feat(continuity): point the next harness at a session cut off without ending”; and 31 more (refs \#1983, \#1165, \#1802).
companions/ai-memory-ts · medium confidence · unverified
497 commits (220 fixes) modifying crates/ai-memory-mcp/src
A change to existing behaviour in crates/ai-memory-mcp/src — 497 commits (220 fixs), 13 files: “feat: allow authenticated users to purge owned sessions”; “fix(wiki-ux): stop machinery from drowning knowledge in the human wiki”; “fix(auto-scope): document and guard session fallback”; and 493 more (refs \#1033, \#387, \#539, \#1253, \#621, …).
crates/ai-memory-mcp/src · medium confidence · unverified
6 commits (3 fixes) modifying crates/ai-memory-cli/src/commands/status
A change to existing behaviour in crates/ai-memory-cli/src/commands/status — 6 commits (3 fixs), 1 file: “fix(cli): prevent server text from injecting terminal controls”; “fix(cli): clarify first-run readiness and improve terminal presentation”; “feat(status): show the client's version and warn across a major version”; and 3 more.
crates/ai-memory-cli/src/commands/status · medium confidence · unverified
7 commits (1 fix) modifying crates/ai-memory-web
A change to existing behaviour in crates/ai-memory-web — 7 commits (1 fix), 3 files: “feat(web): page history with diff and a paged session timeline”; “ai-memory-web: read-only HTTP wiki browser”; “perf(dev): a self-contained build and a two-tier test loop on every platform”; and 4 more.
crates/ai-memory-web · medium confidence · unverified
Introduce local privacy validation and credential sanitation for captured events
The relay now enforces strict privacy checks on event bodies before they are persisted to the local SQLite queue. A new shared \ai-memory-client\ library provides functions to reject sensitive envelopes (containing credentials or unsafe controls) and sanitize external text by redacting secrets like API keys, tokens, and private keys while preserving numeric token metrics. The relay uses these helpers to validate identities, scrub body content, and refuse payloads that attempt to supply destination or actor authority, ensuring that sensitive data is not stored or transmitted. This local sanitation runs in addition to server-side checks, providing a defense-in-depth layer for captured lifecycle events.
companions/ai-memory-relay · high confidence
Introduces per-session and per-actor active project isolation modes
The \active\_project\ module now supports \PerSession\ and \PerActor\ isolation modes (with \PerActor\ as the default) alongside the legacy \Single\ mode. This change allows concurrent agent runs and multiple operators to maintain separate active project contexts without configuration, preventing parallel sessions from overwriting each other's project pointers. It also introduces a \MidSessionRouting\ mode (\Sticky\ vs \FollowCwd\) to control how mid-session events are attributed when the working directory changes.
crates/ai-memory-core/src · high confidence
New 'Management' UI theme with accessible styling
The web interface introduces a new 'Management' visual theme, defined in the static CSS files. This theme applies a specific typographic system (Management Sans, Display, and Mono) and a quiet operational color palette that supports both light and dark modes. It also enforces accessibility standards by ensuring proper contrast ratios and providing visible focus indicators for keyboard navigation. Also in this area: 4 commits (4 fixes) fixing crates/ai-memory-web.
crates/ai-memory-web/static · medium confidence · unverified
Fixes
107 commits (74 fixes) fixing crates/ai-memory-workstream
A fix in crates/ai-memory-workstream — 107 commits (74 fixs), 8 files: “fix(cli): find Claude backfill transcripts by real path”; “feat(run): remember ai-jail choices, mount sibling folders, print the next-run line”; “fix(workstream): skip an oversized transcript record instead of failing the session”; and 101 more (refs \#385, \#820, \#2020, \#1040, \#983, …).
crates/ai-memory-workstream · medium confidence · unverified
17 commits (11 fixes) fixing crates/ai-memory-wiki/src/migrations
A fix in crates/ai-memory-wiki/src/migrations — 17 commits (11 fixs), 3 files: “fix(wiki): skip log ledgers in the OKF nonconformance scan (\#669)”; “feat(okf): backup-gated in-place migration of existing stores”; “fix(wiki): confine page I/O to the project tree”; and 14 more (refs \#669, \#633, \#629, \#698, \#695).
crates/ai-memory-wiki/src/migrations · medium confidence · unverified
220 commits (126 fixes) fixing crates/ai-memory-wiki/src
A fix in crates/ai-memory-wiki/src — 220 commits (126 fixs), 13 files: “fix(wiki): stop reindexing the page of a purged session (\#701)”; “feat: allow authenticated users to purge owned sessions”; “fix(wiki): preserve metadata across reindex”; and 214 more (refs \#701, \#387, \#539, \#669, \#1033, …).
crates/ai-memory-wiki/src · medium confidence · unverified
256 commits (138 fixes) fixing crates/ai-memory-consolidate/src
A fix in crates/ai-memory-consolidate/src — 256 commits (138 fixs), 27 files: “consolidate: tighten anti-hallucination prompt + 4-way provider comparison”; “fix(wiki-ux): stop machinery from drowning knowledge in the human wiki”; “bootstrap: resolve repo root via libgit2, drop git-binary dep”; and 252 more (refs \#1880, \#621, \#937, \#1033, \#884, …).
crates/ai-memory-consolidate/src · medium confidence · unverified
33 commits (26 fixes) fixing crates/ai-memory-core/locales
A fix in crates/ai-memory-core/locales — 33 commits (26 fixs), 10 files: “feat(run): remember ai-jail choices, mount sibling folders, print the next-run line”; “fix(i18n): catalog bootstrap concurrency validation”; “fix(restore-agents): validate the gzip trailer before applying assets”; and 30 more (refs \#2158, \#2020, \#2172, \#2181, \#2119, …).
crates/ai-memory-core/locales · medium confidence · unverified
4 commits (3 fixes) fixing hooks/command-code
A fix in hooks/command-code — 4 commits (3 fixs), 8 files: “fix(hooks): wrap the PowerShell session-start handoff in its harness envelope”; “fix(hooks): send the \[briefing\] keys on the session-start handoff GET (\#998)”; “fix(hooks): never repeat inbox-notice query keys; compaction starts claim nothing”; and 1 more (refs \#998).
hooks/command-code · medium confidence · unverified
42 commits (26 fixes) fixing companions/ai-memory-importer/src
A fix in companions/ai-memory-importer/src — 42 commits (26 fixs), 5 files: “fix(importer): resolve Office relationship namespaces”; “fix(importer): skip fenced code when deriving a page title”; “fix(importer): accept the Open XML SDK worksheet part”; and 39 more (refs \#507).
companions/ai-memory-importer/src · medium confidence · unverified
455 commits (249 fixes) fixing crates/ai-memory-hooks/src
A fix in crates/ai-memory-hooks/src — 455 commits (249 fixs), 19 files: “fix(hooks): answer a refused recap like an unknown session”; “fix(recap): say "1 earlier prompt" when one prompt is omitted”; “fix(hooks): let a linked worktree adopt its main checkout's 2.x project (\#1880)”; and 451 more (refs \#1880, \#1241, \#2012, \#1021, \#1360, …).
crates/ai-memory-hooks/src · medium confidence · unverified
66 commits (42 fixes) fixing hooks
A fix in hooks — 66 commits (42 fixs), 1 file: “fix(hooks): preserve ingest key across shell retry”; “fix(hooks): read a spool entry's fields in linear time (\#727)”; “feat(hooks): script hooks fail closed on a server profile (\#992)”; and 63 more (refs \#727, \#992, \#1033, \#870, \#732, …).
hooks · medium confidence · unverified
7 commits (5 fixes) fixing companions/ai-memory-client
A fix in companions/ai-memory-client — 7 commits (5 fixs), 4 files: “fix(sanitize): redact ai-memory's own credentials”; “fix(client): anchor the companion scrub's sk- rule like the server's (\#2025)”; “fix(sanitize): match credentials across invisible characters and redact more key blocks”; and 4 more (refs \#2025).
companions/ai-memory-client · medium confidence · unverified
8 commits (6 fixes) fixing crates/ai-memory-cli/src/commands/doctor
A fix in crates/ai-memory-cli/src/commands/doctor — 8 commits (6 fixs), 2 files: “fix(cli): prevent server text from injecting terminal controls”; “fix(cli): neutralize the doctor capture coverage line”; “feat(doctor): report Hermes Agent hook consent per profile”; and 5 more (refs \#1745).
crates/ai-memory-cli/src/commands/doctor · medium confidence · unverified
8 commits (6 fixes) fixing hooks/gemini-cli
A fix in hooks/gemini-cli — 8 commits (6 fixs), 28 files: “fix(hooks): send the \[briefing\] keys on the session-start handoff GET (\#998)”; “fix(hooks): drop a Gemini Stop carrying the reply in the script bundles”; “fix(hooks): never repeat inbox-notice query keys; compaction starts claim nothing”; and 5 more (refs \#998, \#386).
hooks/gemini-cli · medium confidence · unverified
9 commits (7 fixes) fixing companions/ai-memory-macos/Sources/AIMemoryMenuCore
A fix in companions/ai-memory-macos/Sources/AIMemoryMenuCore — 9 commits (7 fixs), 9 files: “fix(macos): escape LaunchAgent paths”; “fix(auth): let a browser and the Mac app into a fresh token install”; “docs(macos): trim ServerLog doc comments to the companion's style”; and 6 more.
companions/ai-memory-macos/Sources/AIMemoryMenuCore · medium confidence · unverified
Automated changelog integrity and packaging validation scripts
New scripts enforce release hygiene and packaging correctness: changelog\_sections.py and check-changelog-frozen.sh prevent accidental edits to already-released CHANGELOG sections by comparing versions across the main file and the docs/changelog archive, while check-changelog-sections.sh rejects duplicate headings within version blocks. Packaging validation is handled by check-native-packaging.sh (verifying systemd/sysusers/tmpfiles syntax and ensuring the optional TypeScript SDK is excluded from server distributions), check-nix-packaging.sh (smoke-testing the Nix build output), and check-migrations-immutable.sh (blocking modifications to shipped SQL migrations).
scripts · high confidence
Test coverage
18 commits adding/updating tests in tests/e2e; 24 commits adding/updating tests in crates/ai-memory-hooks/tests/suite; 26 commits adding/updating tests in tests; 4 commits adding/updating tests in evals/tests; 6 commits adding/updating tests in crates/ai-memory-wiki/tests; 7 commits adding/updating tests in crates/ai-memory-cli/tests/fixtures; 8 commits adding/updating tests in companions/ai-memory-macos/Tests; 80 commits adding/updating tests in tests/hooks; Added integration tests for admin and memory access features; Added multi-operator acceptance test harness; Added test coverage for memory consolidation features; Added test coverage for memory store features; Added test suite for ai-memory CLI; Added test support for simulating Codex CLI behavior; Expanded integration test coverage for LLM providers and embedders; Integration tests for the web UI's graph, health, history, and management features.
Dependencies
Workspace restructured and upgraded to version 2.4.2 with new crates and dependency updates
The project has been upgraded to version 2.4.2, introducing several new crates: \ai-memory-web\ (a read-only HTTP wiki browser), \ai-memory-workstream\ (native harness adapters), and \ai-memory-test-support\ (shared test helpers). The workspace now includes \evals\ for live A/B testing. Key dependency updates include \rmcp\ to 2.2.0 for MCP transport security, \rustls\ to 0.23.45 to address RUSTSEC-2026-0285, and \h2\ to 0.4.16 for RUSTSEC-2026-0258. The \reqwest\ client is now configured to use native platform TLS roots instead of bundled webpki roots to support private CAs. Additional dependencies like \clap\_complete\, \crossterm\, \toml\edit\, \jsonc-parser\, \argon2\, \rusqlite\, \refinery\, \notify\, \git2\, and \candle-\\ have been added to support new features such as shell completions, interactive pickers, config file preservation, password hashing, SQLite migrations, file watching, wiki versioning, and local embeddings. Also in this area: 1 commit updating dependencies (3 manifests); 2 commits updating dependencies (2 manifests); Version bump to 3.2.7; Updated project homepage URL; Replace unmaintained serde\_yaml with serde-saphyr and update workspace versions; and 1 more.
(dependencies) · medium confidence · unverified
Written by cca.canine.dev from the codebase's own history, inside the signed delivery this page is composed from.
How this codebase got here
Score
- CAI 65 → 68 (+2.7)
- Rubric changed (rubric-2026.10.5 → rubric-2026.11.4) — scores are not directly comparable.
Lenses
- Code Health 63 → 64 (+1.1)
- Architecture 68 → 69 (+1.0)
- Maturity 85 → 86 (+1.5)
- Readiness 63 → 69 (+6.1)
- Security 69 → 68 (-1.3)
- Accessibility 74 → 79 (+5.6)
- Performance 70 → 100 (+30.0)
Resolved (160)
- Change coupling: cli.rs ↔ lib.rs (crates/ai-memory-cli/src/cli.rs)
- Consolidator::consolidate_session_multi (cognitive 28) (crates/ai-memory-consolidate/src/consolidator.rs)
- Consolidator::consolidate_session_multi (cyclomatic 19) (crates/ai-memory-consolidate/src/consolidator.rs)
- Duplicated block (10 lines × 2) (companions/ai-memory-importer/src/agentmemory/export.rs)
- Duplicated block (10 lines × 2) (crates/ai-memory-cli/src/commands/install_mcp.rs)
- Duplicated block (10 lines × 2) (crates/ai-memory-store/src/ops.rs)
- Duplicated block (10–11 lines × 2) (crates/ai-memory-cli/src/commands/hook_drain_process.rs)
- Duplicated block (10–11 lines × 2) (crates/ai-memory-mcp/src/admin.rs)
- Duplicated block (10–12 lines × 2) (crates/ai-memory-cli/src/config.rs)
- Duplicated block (10–14 lines × 2) (crates/ai-memory-mcp/src/admin.rs)
- Duplicated block (11 lines × 2) (crates/ai-memory-cli/src/commands/auto_improve_report.rs)
- Duplicated block (11 lines × 2) (crates/ai-memory-cli/src/commands/doctor.rs)
- Duplicated block (11 lines × 2) (crates/ai-memory-cli/src/commands/install_hooks.rs)
- Duplicated block (11 lines × 2) (crates/ai-memory-mcp/src/admin.rs)
- Duplicated block (11 lines × 2) (crates/ai-memory-store/src/reader.rs)
- Duplicated block (11–14 lines × 2) (crates/ai-memory-cli/src/commands/move_project.rs)
- Duplicated block (12 lines × 2) (crates/ai-memory-cli/src/commands/install_hooks.rs)
- Duplicated block (12 lines × 2) (crates/ai-memory-store/src/reader.rs)
- Duplicated block (12 lines × 2) (crates/ai-memory-store/src/reader.rs)
- Duplicated block (12 lines × 2) (crates/ai-memory-workstream/src/transcript.rs)
- …and 140 more
New (715)
- AiMemoryClient.query (cyclomatic 17) (companions/ai-memory-ts/src/client.ts)
- AiMemoryServer::memory_write_pages_batch (cognitive 29) (crates/ai-memory-mcp/src/server.rs)
- AiMemoryServer::memory_write_pages_batch (cyclomatic 19) (crates/ai-memory-mcp/src/server.rs)
- AmbiguousProjectHolders::fmt (cognitive 17) (crates/ai-memory-store/src/error.rs)
- Change coupling: client.ts ↔ coverage.ts (companions/ai-memory-ts/src/client.ts)
- Change coupling: coverage.ts ↔ types.ts (companions/ai-memory-ts/src/coverage.ts)
- Change coupling: curator.rs ↔ writer.rs (crates/ai-memory-consolidate/src/curator.rs)
- ClaudeCodeTotals::add (cognitive 21) (crates/ai-memory-cli/src/commands/session_usage.rs)
- Consolidator::consolidate_session_multi_in_scope (cognitive 34) (crates/ai-memory-consolidate/src/consolidator.rs)
- Consolidator::consolidate_session_multi_in_scope (cyclomatic 22) (crates/ai-memory-consolidate/src/consolidator.rs)
- Critical CVE: [GHSA redacted] (Cargo.lock)
- Duplicated block (10 lines × 2) (companions/ai-memory-importer/src/agentmemory/export.rs)
- Duplicated block (10 lines × 2) (companions/ai-memory-importer/src/mcp.rs)
- Duplicated block (10 lines × 2) (crates/ai-memory-cli/src/commands/checkpoints.rs)
- Duplicated block (10 lines × 2) (crates/ai-memory-cli/src/commands/install_mcp.rs)
- Duplicated block (10 lines × 2) (crates/ai-memory-cli/src/commands/install_mcp.rs)
- Duplicated block (10 lines × 2) (crates/ai-memory-cli/src/commands/move_project.rs)
- Duplicated block (10 lines × 2) (crates/ai-memory-consolidate/src/dream.rs)
- Duplicated block (10 lines × 2) (crates/ai-memory-llm/src/redirect.rs)
- Duplicated block (10 lines × 2) (crates/ai-memory-store/src/reader.rs)
- …and 695 more
Changes since last survey
- 300 commits — 167 feature/other, 133 fixes
By area
- (repo) — 105 commits
- (root) — 94 commits
- crates/ai-memory-cli — 23 commits
- crates/ai-memory-hooks — 20 commits
- crates/ai-memory-store — 12 commits
- crates/ai-memory-mcp — 9 commits
- crates/ai-memory-core — 5 commits
- crates/ai-memory-web — 5 commits
- docs/security-boundaries.md — 4 commits
- companions/ai-memory-importer — 3 commits
- crates/ai-memory-llm — 3 commits
- companions/ai-memory-ts — 2 commits
- crates/ai-memory-consolidate — 2 commits
- hooks/gemini-cli — 2 commits
- companions/ai-memory-macos — 1 commit
- companions/ai-memory-relay — 1 commit
- companions/ai-memory-wikisync — 1 commit
- docs/https-via-proxy.md — 1 commit
- docs/install.md — 1 commit
- docs/lifecycle-ops.md — 1 commit
Notable commits
- fix: Fix profile forget/show path handling and honest delete reporting (#2181)
- fix: Merge branch 'main' into fix/gemini-token-budget-stop
- fix: Merge branch 'main' into fix/health-unreadable-storage-not-zero
- fix: Merge branch 'main' into fix/install-mcp-verbatim-data-dir
- fix: Merge branch 'main' into fix/web-envelope-svg-icon
- fix: Merge pull request #2197 from matheus-rodrigues00/fix/2195-drop-subagent-keeps-main
- fix: Merge pull request #2198 from matheus-rodrigues00/fix/2196-gemini-prompt-capture
- fix: Merge pull request #2199 from cristianob/fix/profile-truncated-batch-split
- fix: Merge pull request #2201 from felipe-NR/fix/2200-claude-resume-session-link
- fix: Merge pull request #2205 from fabiomotta0311/fix/install-mcp-verbatim-data-dir
- fix: Merge pull request #2209 from samirhvbr/fix/restore-validate-archived-config
- fix: Merge pull request #2210 from samirhvbr/fix/upgrade-preserve-claude-hook-url
- fix: Merge pull request #2211 from zch-soft/fix/hermes-home-resolution
- fix: Merge pull request #2215 from rudycelekli/fix/health-unreadable-storage-not-zero
- fix: Merge pull request #2216 from rudycelekli/fix/adopt-identity-invalidates-scope-cache
- fix: Merge pull request #2217 from samirhvbr/fix/sdk-literal-percent-path
- fix: Merge pull request #2218 from samirhvbr/fix/gemini-token-budget-stop
- fix: Merge pull request #2219 from samirhvbr/fix/opencode-responses-token-budget-stop
- fix: Merge pull request #2220 from samirhvbr/fix/eval-comparison-cutoff-grids
- fix: Merge pull request #2221 from rudycelekli/fix/rename-project-invalidates-scope-cache
- …and 280 more
Written by cca.canine.dev from the codebase's own history, inside the signed delivery this page is composed from.
Survey your own repository
akitaonrails/ai-memory was measured the same way every project in this corpus was: the same rubric, at a pinned commit, with the result published in full. Point a surveyor at a repository you know and see whether you agree with it.
About this page
- The score is its most recent published measurement, taken on 14 November 2026 at a pinned commit. It is not a live figure and does not change until the project is measured again.
- Measured at commit 73f6ccd036b8490b1072877f76740d96509c7d7a — the exact code this score is about.
- Scored under rubric-2026.11.4 — the same rubric and the same method as every other entry in this index.
- Measured by Canine Code Analyzer (cca.canine.dev) using cca measure-2026.11.65+preprod-e1a3b69fc8ee.