Skip to content

CHANGELOG.md

Latest commit

 

History

History
1033 lines (734 loc) · 72.7 KB

CHANGELOG.md

File metadata and controls

1033 lines (734 loc) · 72.7 KB
Raw

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

Unreleased

  • 3f90d45 Added the built-in dev-retro skill: an evidence-based retrospective that turns friction into durable fixes, automation first. dev-lifecycle and dev-debug offer it and run it only if the user agrees.
  • fdbfb3a Wired dev-smoke into other skills: verify requires it for "feature works" claims on user-facing changes, dev-refactor and dev-simplify compare runs before and after, dev-debug uses it as the repro, and dev-lifecycle runs it before code review.
  • aa06cf5 Added server start, readiness, and teardown helpers plus HTTP status checks to the dev-smoke shell template, and API checks now cover response bodies and error cases.
  • 67dfb01 Extended dev-smoke to UI and web changes: real-browser checks with a disposable Playwright template, fresh browser contexts, console-error checks, and screenshots as evidence.
  • 167a682 Added the built-in dev-smoke skill: it maps the user flows a change can affect, then builds and runs a disposable sandboxed script to smoke-test them end to end.
  • 5dc438c Made agent-orchestration a built-in skill, and skills now refer to other skills by plain name instead of the Codex-only $skill form.
  • 29155fd Renamed the structured-debug skill to dev-debug (previously debug); reinstall it or update .ai-devkit.json to use the new name.
  • 08f1a74 Renamed the simplify-implementation skill to dev-simplify and refactor to dev-refactor, and made dev-refactor a built-in skill; reinstall or update .ai-devkit.json to use the new names.

[0.71.0] - 2026-10-23

  • dad3de5 Migrated interactive agent start into the devkitd daemon via a new agent.start RPC (tmux and herdr runtimes); daemon.db is now the agent registry, and @ai-devkit/daemon-client adds startAgent() and structured DaemonRpcError.
  • 007cbd0 Added multi-agent orchestration, troubleshooting, and permissions docs, and fixed README setup, skill count, and harness table.
  • ff51a61 Fixed agent send --wait failing with "No session file found" right after agent start by waiting for the harness session file to appear.
  • 31f5bb4 Sharpened memory MCP tool descriptions so agents search before work and after errors, and store/update after tasks without duplicates.
  • 05aa14e Aligned web content with shipped features.
  • cea04b3 Added light and dark themes with a header toggle to the website.
  • 162caaf Added web pages for the four missing built-in skills.
  • fec8289 Redesigned the web skills catalog and registry browser.
  • 171c374 Corrected capitalisation of FAQ topic titles.
  • 52eec02 Redesigned the FAQ on the docs reading layout.
  • 6581f6d Redesigned the docs with a grouped sidebar, outline, and search.
  • 9eae9bf Fixed the console showing a stale name after renaming an agent; the list now refreshes and reselects the renamed agent.
  • 66d2ea6 Updated web content.

[0.70.1] - 2026-10-12

  • 70e95fc Expanded the web FAQ, SEO keywords, and social preview image.
  • 2475c07 Redesigned the homepage with Remotion demos and accessibility fixes.
  • f3379b8 Added Grok CLI and Muse Code to the supported-agents page.
  • 033f1f1 Corrected all license references to Apache-2.0: the LICENSE file has always been Apache License 2.0, but package.json fields, Cargo.toml, plugin manifests, READMEs, and web content incorrectly stated MIT.
  • cdb15b6 Added the full-logo navbar assets.
  • d348283 Rendered the full logo in the navbar.
  • de852a6 Moved Telegram bot token validation into channel-connector's shared validateTelegramBotToken, redacting the bot token from debug output.
  • 86d46ac Fixed devkitd parsing ps rows with padded columns on Linux, which left agent list permanently empty; empty daemon answers are now cross-checked against local detection.

[0.70.0] - 2026-10-10

  • 5f8c4a4 Fixed stale agent types, durable modes, and supported tools in the docs.
  • df6fb0c Fixed stale memory MCP setup instructions to match the setup-first wiring.
  • 20ab9a9 Rewrote the console start-agent pane docs for the 0.68 UX.
  • 7611135 Documented agent session compact.
  • 21e6d54 Added session-compact to the built-in skills docs.
  • 04a9972 Documented memory semantic and memory reembed commands.
  • 72746bf Mentioned the Herdr managed runtime in the docs.
  • b1f4b79 Documented agent send --stdin, --wait, --group, and --json.
  • 96204cb Generalized the channel agent prerequisite to any detected harness.
  • 1ce388f Updated the getting-started setup description and first-run checks.
  • 4b513a9 Removed the duplicated prerequisites block.
  • 16b6ca2 Refreshed the supported-agents example and contributor path.
  • 1ecb476 Added the observe-and-operate capability to the overview.
  • 651f080 Updated the AI lifecycle docs indexes.
  • d0c42a0 Added Muse Code skill install support for project and global installs.
  • 708897d Added the Muse Code harness across agent-manager, the CLI, and the Rust devkitd daemon.
  • 31047e1 Gated fmt and lint checks for Rust and TypeScript in CI.

[0.69.1] - 2026-10-09

  • c085143 Fixed the bundled agent console crashing on load: esbuild code splitting exposes only the default export on the lazy React chunk, so destructured createElement was undefined in the published bundle.
  • 3b60a74 Fixed the audit configuration.

[0.69.0] - 2026-10-09

  • Introduced devkitd, a Rust coordination daemon with auto-spawning unix-socket server, harness ports for every agent type, per-platform binary packages, and a daemon command group in the CLI (#352).

[0.68.0] - 2026-10-08

  • Wired the memory MCP server globally into all MCP-capable harnesses during setup (#308).
  • Reported Devin usage in capacity (#348).
  • Documented antigravity_cli and Devin in the start-types table (#347).
  • Rendered the start-agent type selector as a vertical list (#329).
  • Supported j/k field navigation in the start-agent pane (#331).
  • Submitted the start-agent form on Enter in the name field (#334).
  • Showed multiline error output in the start-agent pane (#336).
  • Made Ctrl+C close the active pane before quitting (#337).
  • Selected the newly started agent on success (#344).
  • Remembered the last-used type and cwd across start-agent pane opens (#345).
  • Showed friendly agent type labels in the start-agent pane (#330).
  • Supported Shift+Tab reverse focus in the start-agent pane (#332).
  • Allowed cancelling a pending agent start and showed elapsed time (#335).
  • Showed contextual key hints per focused field (#333).
  • Mapped q to cancel in the start-agent pane (#338).
  • Exposed interactive/durable mode in the start-agent pane (#339).
  • Added an optional initial prompt field to the start-agent pane (#340).
  • Pre-checked harness availability in the start-agent pane (#341).
  • Validated name and cwd client-side before submit (#342).
  • Expanded ~ in cwd and offered recent project dirs (#343).
  • Added a passthrough args field to the start-agent pane (#346).
  • Split agent commands into modules and consolidated shared helpers with no public API changes (#350).

[0.67.0] - 2026-10-05

  • Added the Devin CLI harness adapter (#309).
  • Ranked skill find results and showed install commands (be4e7be).
  • Added --mode to skill add (24427c6).
  • Configured per-skill install mode with copy or link (362569c).
  • Documented WezTerm and Ghostty as supported terminals (0534cd3).
  • Added a Ghostty terminal backend (a5cf236).

[0.66.0] - 2026-09-30

  • Started the memory MCP server when launched through the bin symlink (#303).
  • Showed Anthropic usage for the Pi harness (708e12c).
  • Separated harness credentials from capacity providers (cf1319e).
  • Simplified the agent-manager runtime and terminal structure (bd76b05).
  • Colocated readiness checks with their harnesses (817302d).
  • Reported readiness for all harnesses (3c0666f).

[0.65.1] - 2026-09-29

  • Cached incremental session summaries across refreshes so repeated agent listings skip rescanning unchanged transcripts (#267).
  • Used a shared process snapshot in the Antigravity adapter (#268).
  • Excluded Codex helper processes from agent detection so helper invocations no longer show up as agents (#270).
  • Read conversation tails from the end of JSONL transcripts instead of whole files when rendering conversation previews (#271).
  • Read only the session metadata head during Codex session discovery (#269).
  • Bounded Gemini session discovery I/O (#274).
  • Detected Gemini session-*.jsonl logs (#284).
  • Bounded the cold-start session summary to a head and tail scan (#272).
  • Added skip hooks to the Pi, Grok, and Copilot reducers so bounded scans can skip irrelevant transcripts (#285).
  • Classified Codex review, cloud, and agents sessions by rollout behaviour (#289).
  • Enriched only candidate PIDs and cached process start times during detection (#277).
  • Detected executable paths containing spaces (#286).
  • Shared process/session matching tolerance across locators (#288).
  • Folded the first tail line correctly when the bounded scan starts on a line boundary (#299).
  • Resolved spaced argv[0] values in AGENTS matchers (#300).
  • Consumed variadic Codex -i/--image values in helper detection (#301).
  • Moved the Antigravity adapter into the agent-manager provider structure (29ee399).
  • Moved the Kiro adapter into the agent-manager provider structure (0ae18de).
  • Followed Kiro's bundled bun TUI when matching session locks (6788bd7).
  • Renamed the internal providers structure to harnesses with no public API changes (dd31803).
  • Registered built-in harnesses from one list, exporting createBuiltinAdapters and AGENT_TYPES (8f0a3c5).
  • Shared harness boilerplate in harnesses/shared (b60631a).
  • Added findSessionsById lookups for Kiro and Antigravity sessions so exact historical IDs resolve without enumerating summaries (cc43cbe).
  • Derived StartableAgentType from AGENT_TYPES (cd8a7e7).
  • Cached Kiro and Antigravity session summaries (532c562).
  • Released @ai-devkit/agent-manager 0.37.0 with this version, exporting AGENT_TYPES, createBuiltinAdapters, and the ConversationOptions type; the providers → harnesses rename is internal-only with no public API changes.

[0.65.0] - 2026-09-27

  • Added an OpenAI platform usage provider to ai-devkit capacity, resolving OPENAI_API_KEY or the Pi OpenAI credential and reporting today and 7-day token windows from the platform Usage API; unqueryable limits render consumption-only (#252).
  • Added Claude subscription usage to ai-devkit capacity, reading quota windows from the local Claude login (#251).
  • ai-devkit capacity now shows OpenAI usage from pi's OpenAI login alongside harness quotas (#254).
  • Added the Kiro CLI adapter — detects running Kiro CLI agents by matching ~/.kiro/sessions/cli lock files to session metadata and transcripts, enabling Kiro in agent list/detail and channels (#115).
  • Added the Antigravity CLI adapter — detect, list, and inspect Google Antigravity CLI (agy) agents alongside Claude, Codex, Gemini, and Grok (#133).
  • Resolved agent session detail by exact session ID instead of enumerating every provider's history (~5.8s → ~0.45s with ~1.3 GB of Codex sessions), with identical output (e8a93ea).
  • Released @ai-devkit/agent-manager 0.36.0 with this version, exporting KiroAdapter, AntigravityCliAdapter, and the new getClaudeCapacityReport/getOpenAiCapacityReport probes for programmatic use.

[0.64.0] - 2026-09-24

  • Added ai-devkit agent session compact — Jev-backed compaction of a historical session into a structured continuation artifact (durable continuation facts, memory candidates), with a new session-compact built-in skill (#250).
  • Added AgentManager.findSessionsById() and the optional AgentAdapter.findSessionsById() hook to @ai-devkit/agent-manager, resolving exact historical session IDs across providers without enumerating every summary (9684956).
  • Moved the grok adapter into the agent-manager provider structure with no public API changes (3256cc3) via @ai-devkit/agent-manager 0.35.0 (released with this version).

[0.63.0] - 2026-09-23

  • Redesigned the capacity command output into a single per-quota table with usage bars, threshold colors, humanized reset times (in 4h 40m · 20:25), and used/total amounts instead of duplicated window summaries with truncated ISO timestamps (3c65af7).
  • Split capacity identity into harness and provider (codex | OpenAI, pi | z.ai) via a new harness field on CapacityReport; the Codex probe now reports provider: "openai" (3c65af7).
  • ai-devkit capacity without arguments now probes every supported provider and renders one Harness/Provider/Status table; --json emits a single object for one provider and an array for many, and per-provider probe failures warn instead of aborting the whole listing (3c65af7).
  • Added global z.ai usage limits to ai-devkit capacity through a new getZaiCapacityReport probe in @ai-devkit/agent-manager (4ea4868).
  • d58171c Added shared UI formatting utilities (pluralization, status labels, time formatting) reused across CLI commands.
  • 20e909b Clarified agent list and detail rendering.
  • e5043dd Polished channel status rendering.
  • c8d29ba Aligned status output with the visible checks.
  • 6103e40 Refreshed lint human output.
  • 0383f02 Aligned skill result rendering.
  • 325ce86 Capitalized memory --table headers and shortened the UUID column.
  • f6da095 Standardized flow success output.
  • Moved the opencode adapter into the agent-manager provider structure with no public API changes (f34a9a3) via @ai-devkit/agent-manager 0.34.0 (released with this version).

[0.62.2] - 2026-09-14

  • 7af9864 Fixed agent status losing registry routing when the agent list is refreshed, and cleaned up opencode process listeners on stop (1a3a002) via @ai-devkit/agent-manager 0.33.1 (pinned in this release; also moves the copilot and gemini adapters into the provider structure with no public API changes).
  • Re-pinned @ai-devkit/memory to 0.20.1 and @ai-devkit/channel-connector to 0.13.4 — parity patch releases with no functional changes (tooling/format only).
  • Published previously unreleased package versions and added their missing publish workflows: @ai-devkit/task-manager 0.4.0 (WAL concurrency fix, #206), @ai-devkit/memory-dashboard 0.1.3, @ai-devkit/pi-session-tracker 0.1.1.
  • Clarified the Telegram channel authorization flow in the remote skill docs (26fff09) and added a release smoke-test script plus a Codex plugin build script.

[0.62.1] - 2026-09-11

  • Fixed the published package composition: 0.62.0 shipped importing AGENT_RUNTIME_PROVIDERS and other exports from @ai-devkit/agent-manager 0.33.0 while pinning 0.32.0, breaking fresh installs. 0.62.1 pins @ai-devkit/agent-manager 0.33.0 (published with this release). 0.62.0 has been removed from npm.

[0.62.0] - 2026-09-11 — yanked (broken dependency pin)

  • 5941ca5 Added local folder skill registries.
  • c661165 Reorganized CLI skill services.
  • dafaedb Moved the skill UI out of services.
  • 45b76a9 Moved the Codex adapter into the agent-manager provider structure.
  • 1a858a4 Moved the Pi adapter into the agent-manager provider structure.
  • b4d2b8f Updated the license copyright year and owner.
  • 033c4ef Added the Herdr agent runtime.

[0.61.0] - 2026-09-05

  • 7ad2afb Show tmux guidance after setup output as next steps.
  • 1547913 Fetch the built-in skills list from a remote manifest so new built-ins can ship without a CLI release, with an embedded fallback for offline use.

[0.60.0] - 2026-09-03

  • d105b57 Reduced hybrid semantic noise with a 0.50 cosine floor for semantic-only results and a tighter RRF constant (60 to 10), preserving the full recall gain: judged irrelevant top-3 results fell from 4.7% to 2.5%, known-bad results from 6 to 3, and hit@3 remained 97%.

[0.59.0] - 2026-09-02

  • f077fa1 Added a read-only tmux prerequisite check to setup and platform-aware installation hints to managed agent startup.
  • 20d799f Reorganized the memory package into domain, repository, semantic, and service layers.
  • 8e65a5d Added optional You.com search integration to the skill registry.
  • 0d14c87 Added adapter readiness checks to agent status reporting.

[0.58.0] - 2026-09-01

  • 6d4fdc1 Added opt-in hybrid semantic memory search with embedding storage, maintenance, and CLI configuration.

[0.57.1] - 2026-08-31

  • Pinned @ai-devkit/memory to 0.17.0 so CLI installs include the database WAL concurrency fix (#206).

[0.57.0] - 2026-08-31

  • 0f2d731 Made WAL setup concurrency-safe across packages so concurrent memory store and CLI processes no longer hit database-locked errors at startup.

[0.56.0] - 2026-08-28

  • 0039ac7 Added the MIT license.
  • 7d66b48 Aligned onboarding docs with the setup-then-init flow.
  • 2b7dd5e Added pi print mode.
  • 06bb8b9 Prepared skill registries once per CLI run.
  • 4103bd5 Completed the init flow with a shared project application service.

[0.55.0] - 2026-08-23

  • 7f33392 Removed the skill registry.
  • 7dbcc09 Added the provider capacity command.

[0.54.0] - 2026-08-21

  • 8284e76 Indexed cached registry skills.

[0.53.0] - 2026-08-21

  • 4ac8fa2 Expanded the agent CLI web reference.
  • 5192336 Documented console controls and agent types.
  • 5ac70fb Aligned the built-in skill catalog.
  • b5a2bce Corrected the structured debug skill name in docs.
  • 74ff63d Expanded send wait adapter coverage.
  • 4c6ec2c Clarified channel provider behavior.
  • 4233e1c Clarified init overwrite behavior.
  • c8a1fee Listed first-party plugins in docs.
  • 4b15897 Refreshed the skill management roadmap.
  • 870529b Completed the agent management roadmap.
  • 0d85d5b Updated hooks and memory roadmaps.
  • c6c2c70 Surfaced durable task tracking in docs.
  • 4f3afbf Generalized console channel guidance.
  • 70becde Refined the CLI enhancements roadmap.
  • db7a762 Added the refactor skill.
  • 44f47ef Tightened refactor skill guidance.
  • 43e2ca9 Simplified the refactor skill.
  • 4ee608c Stored durable agents in agents.db.
  • 95674c9 Reorganized channel provider adapters.
  • dabf19f Renamed the durable agent start mode.
  • 8c1e490 Preferred current Codex sessions in the agent manager.
  • 9dced35 Added a scroll-safe working animation to the running agent's console preview, with a TERM=dumb ASCII fallback.

[0.52.0] - 2026-08-18

  • 4690e3e Simplified obsolete test cleanup guidance in the simplify-implementation skill.
  • 7b711bd Documented building workspace artifacts before full gates and commit hooks.
  • b9c0603 Rendered inline HTML inside Telegram list items.
  • 363682f Added agent pinning in the console list.
  • 1f62bfc Added agent name filtering in the console.

[0.51.0] - 2026-08-16

  • 5d9dc3f Rendered Markdown in the agent console preview pane.
  • 1ad381b Updated package audit lockfile metadata.

[0.50.1] - 2026-08-15

  • b3dd3b9 Treated only ESRCH errors as definitive process death in the agent manager.
  • bea4e9d Reduced agent registry refresh writes.

[0.50.0] - 2026-08-14

  • ab2a6a3 Added agent modes to CLI agent listings.
  • 3dc4ea6 Kept CLI agent list rows on one line.
  • ef0184f Memoized console preview rows to improve rendering performance.
  • 8864553 Isolated console message input state to reduce rerenders.
  • d84a53c Split console state into focused contexts.
  • 051b941 Shared asynchronous process snapshots in the agent manager.

[0.49.1] - 2026-08-14

  • ad40b7e Clarified agent console preview formatting and improved rendering performance.
  • 7787f3e Stored the agent registry in SQLite.

[0.49.0] - 2026-08-13

  • 9593d83 Fixed Codex session message parsing in the agent manager.
  • a5a124f Added the skill add-registry CLI command.

[0.48.0] - 2026-08-12

  • 122d8dd Simplified the task manager README.
  • 7628b9e Added task names to tracing.
  • 1ed5d9a Added memory list display to the agent console.
  • 4524973 Added a focusable scrollable detail/chat pane with the v shortcut.
  • 15b09d2 Fixed invalid YAML frontmatter in the agent-orchestration skill metadata.
  • 5bfd467 Added the brainstorm skill.
  • 7ee6def Fixed init template composition with built-in skills.
  • 85fb872 Fixed agent detection at tmux pane PID.
  • 2126c04 Added CLI listing for globally installed skills.
  • f231c06 Added support for removing globally installed skills.
  • 2c01e03 Added the Slack Socket Mode channel connector.
  • 72a6c37 Added durable Claude print mode for agents.

[0.47.0] - 2026-07-02

  • 941fabf Added Grok environment skills support.
  • 4206681 Added the Grok Build CLI adapter.
  • aa972da Added Antigravity CLI environment support.
  • fbf4bf7 Added the durable task system with append-only event history.
  • 8ecb865 Added the built-in task skill for CLI-driven lifecycle/debug progress tracking.
  • 5c69c3a Updated better-sqlite3 to 12.11.1.

[0.46.0] - 2026-06-30

  • 172ef94 Added Claude PreToolUse forwarding to Telegram for tool invocations.
  • 61a6633 Updated the Claude Code hook configuration.
  • 7fa1ff Added WezTerm terminal support to agent manager.
  • 70c86c9 Rendered Claude AskUserQuestion prompts as Telegram inline keyboards.

[0.44.1] - 2026-06-29

  • fb66494 Fixed tmux send body delivery by using bracketed paste.
  • d75ab3b Fixed Telegram markdown delivery by chunking messages before rendering.

[0.44.0] - 2026-06-27

  • f7d1144 Fixed Telegram delivery reliability.
  • e638fdf Removed updatedAt from new project config writes.
  • 291ffdf Removed the install artifact overwrite prompt.
  • 76887e6 Added the dev-commit workflow skill.
  • 05581dd Added the dev-pr workflow skill.
  • e04ae44 Installed built-in skills during agent setup.

[0.43.0] - 2026-06-26

  • 997ef2d Added the agent-management built-in skill.
  • 0b1cc33 Added the Codex session mapping hook.
  • 3b5b99c Added the agent setup CLI command.

[0.42.1] - 2026-06-25

  • 8e5e745 Fixed Claude agent start timeout handling.

[0.42.0] - 2026-06-17

  • 676bd04 Fixed dev-lifecycle startup guidance to list installed skills.
  • 60f9270 Improved the simplify-implementation skill guidance.
  • ad06c60 Added the memory dashboard plugin.
  • 34d3552 Added CLI support for agent groups.
  • e734194 Fixed Codex session matching by using session_meta timestamps.

[0.41.2] - 2026-06-14

  • c75d73b Split dev-lifecycle into orchestrated phase-specific skills.
  • 12ba398 Added AI DevKit branding to skill metadata.
  • 2deeb57 Updated built-in skill definitions.

[0.41.1] - 2026-06-13

  • b0c3d45 Documented Pi agent integration.
  • bd40286 Fixed duplicate Copilot wrapper agents.
  • 8a943ff Fixed wrapper PID handling for Gemini and Copilot agents.

[0.41.0] - 2026-06-13

  • ab88047 Added package repository metadata.
  • de3203b Added Junie CLI support.
  • 807ec44 Added Cline support.
  • 1520d2f Updated GitHub Copilot support.
  • aeb8a15 Added Devin support.
  • 432b32f Removed unused environment context filename metadata.
  • 8c1bf9e Removed workflow slash command support.
  • 39562d2 Added Gemini project skill path support.
  • aa43342 Added Roo Code skills and MCP support.
  • 672496a Removed Windsurf environment support.
  • fbaef04 Added Kilo Code skills and MCP support.
  • ce437db Added Pi environment skills support.
  • 36e4b29 Added OpenCode MCP support.
  • b4bdabd Fixed macOS terminal app process detection in agent manager.
  • 8dc7f98 Added Pi support to agent start.
  • 77c0413 Added Copilot support to agent start.

[0.40.0] - 2024-06-11

  • f0f72d0 Added the Pi session tracker extension.
  • 57b08bc Added the Pi adapter to agent manager.
  • 69fb18a Optimized Codex resume session lookup.
  • 35d4dee Added the global plugin foundation to the CLI.

[0.39.0] - 2024-06-09

  • f9bb55c Reverted the CLI startup optimization from #96.
  • 74b3139 Added agent rename support inside the agent console.
  • 1884f56 Upgraded project dependencies.
  • fcff530 Migrated CLI prompts to @inquirer/prompts.
  • ad1ad29 Migrated package builds to SWC.
  • c6886f5 Fixed duplicate Gemini wrapper processes in agent discovery.
  • 0038cc5 Added channel controls to the agent console.
  • 2314ac9 Added the agent-communication skill.
  • 1dfe1fe Updated documentation for the channel daemon option.
  • b596aca Added the agent console guide.
  • 9a6d2b0 Refreshed project documentation.
  • 0e746d8 Fixed resumed Codex session matching.
  • e795edf Added a Copilot adapter to agent manager.
  • ec2ca1e Added the changelog skill.

[0.38.0] - 2026-06-02

Added

  • Agent Console Help Pane - Added a help pane to ai-devkit agent console with context-aware keyboard shortcuts and layout support for toggling help while preserving existing console workflows.
  • TUI Design System - Added shared TUI design-system components and tokens for panels, section titles, key hints, spacing, borders, and colors, and updated the agent console to use them consistently.
  • CLI Startup Benchmark - Added a CLI startup benchmark utility, benchmark tests, and maintainer documentation for validating top-level command startup performance.

Changed

  • CLI Startup Performance - Optimized ai-devkit startup by handling version/help output through a lightweight command manifest and lazy-loading Commander plus selected command modules only when needed (#96).

[0.37.0] - 2026-06-01

Added

  • Agent Console Start - Added an in-console start-agent workspace to ai-devkit agent console, allowing users to press s, choose an agent type, edit the generated name and working directory, start the agent through agent start, and refresh the agent list without leaving the TUI (#93).
  • Agent Console Kill - Added agent kill <name> and uppercase K support in ai-devkit agent console with a confirmation dialog, process termination, managed tmux session cleanup, and refreshed console state after successful kills (#94).

[0.36.0] - 2026-05-31

Added

  • Managed Agent Start - Added ai-devkit agent start to launch named agents in managed tmux sessions with --type, --name, and --cwd options, PID polling, stale session cleanup, and registry tracking (#90).
  • Agent Rename - Added ai-devkit agent rename <current-name> <new-name> to rename live registry entries with validation and conflict handling (#92).

Changed

  • Agent Registry Synchronization - Agent discovery now mirrors live Codex and Gemini CLI sessions into the registry so running agents stay available for list/detail/send workflows even when started outside agent start (#91).
  • Agent List Project Column - Updated ai-devkit agent list to show the project name instead of the full working directory for a more compact running-agent table.
  • Dev Lifecycle Guidance - Enhanced the dev-lifecycle skill's test-planning and implementation-log guidance.

Fixed

  • Claude Code Status Detection - Fixed Claude Code sessions being reported with unknown status by improving session parsing for meaningful conversation entries and interrupted requests.

[0.35.0] - 2026-05-28

Added

  • Agent Console - New ai-devkit agent console command that launches a live multi-agent terminal UI (TUI) with an agent list pane, conversation preview, chat input, and status footer for monitoring and interacting with multiple running agents simultaneously (#88).

Changed

  • ESM Migration - Converted all packages (cli, agent-manager, channel-connector, memory) from CommonJS to ES Modules, migrated the test runner from Jest to Vitest, and upgraded project dependencies (#87).
  • Dev Lifecycle Clarification Contract - Tightened the clarification requirements in the dev-lifecycle skill's SKILL.md and updated new-requirement, review-design, and review-requirements references; improved the check-status.sh script.

[0.34.0] - 2026-05-25

Added

  • Added ai-devkit docs init-feature <name> to initialize date-prefixed feature docs for configured phases using the current local date, with optional --json output for agents and scripts.
  • Added lint support for date-prefixed feature docs while preserving compatibility with legacy feature-{name}.md docs.

Changed

  • Updated lifecycle command templates and the dev-lifecycle skill to initialize feature docs through docs init-feature and reference latest matching YYYY-MM-DD-feature-{name}.md docs.

[0.33.0] - 2026-05-24

Added

  • ai-devkit agent session detail now shows detailed historical session information and conversation history, with --json, --type, --full, --tail, and --verbose options.
  • ai-devkit channel connect telegram --name <name> now supports multiple named Telegram channel configurations, including duplicate-token protection and persisted authorized chat IDs.
  • ai-devkit channel start <name> --agent <name> --daemon can run channel bridges in the background, with bridge process tracking, log paths, status reporting, and ai-devkit channel stop [name] support.

Changed

  • ai-devkit channel list and ai-devkit channel status now show authorization and live bridge state for configured channels.

[0.32.0] - 2026-05-22

Added

  • ai-devkit agent send --wait now supports --stdin and piped input, allowing automation scripts to pass prompts without a positional message argument.
  • ai-devkit agent send --wait now supports --timeout <milliseconds> to configure the maximum wait time for an agent response.
  • ai-devkit agent send --wait --json now emits structured JSON with target metadata, the prompt, captured response messages, elapsed time, and final status.
  • Added documentation and FAQ content for automating programmatic agent calls with agent send --wait.

Changed

  • Refreshed README, package descriptions, and web documentation copy to better explain AI DevKit workflows, supported agents, memory, channels, and agent-manager usage.

[0.31.0] - 2026-05-16

Added

  • ai-devkit agent send now accepts --wait to block until the target agent returns a new assistant response, polling the session transcript and printing the captured output on stdout while keeping status/warning logs on stderr (#83).

[0.30.0] - 2026-05-14

Added

  • OpenCode Agent Adapter - Added OpenCodeAdapter to agent-manager so ai-devkit agent commands can discover, inspect, and control running OpenCode sessions.
  • HTML Artifact in document-code Skill - The document-code skill now offers an HTML artifact alongside its markdown output.

Fixed

  • Claude Code PID-File Live Status - agent-manager now prefers the PID-file live status when resolving Claude Code agent state, improving accuracy of agent list output.
  • Claude Code Lossy Project Dir Encoding - agent-manager now matches Claude Code's lossy project directory encoding when resolving session paths, fixing session discovery for paths that Claude Code re-encodes.

[0.29.0] - 2026-05-11

Added

  • Non-interactive init - ai-devkit init now accepts -y, --yes to run without prompts (required for agent/CI contexts where stdin is not a TTY). Without a template, --yes requires -e <env> and one of -a/-p, otherwise it exits non-zero with a clear message instead of hanging on a checkbox prompt.
  • init --overwrite flag - When combined with --yes, --overwrite overwrites existing environments and phase files; the default under --yes is to skip them, matching the install --overwrite convention.
  • Telegram Markdown Rendering - channel-connector's Telegram adapter now renders Markdown to Telegram-flavored HTML with an automatic plain-text fallback when Telegram rejects the formatted payload.
  • Channel Polling Diagnostics - The channel command now surfaces polling-loop errors and emits diagnostic logs to help debug long-running channel sessions.

Changed

  • Dev Lifecycle Bootstrap - The dev-lifecycle skill's prerequisite step now invokes ai-devkit init -a -e claude --built-in --yes so agents (e.g., OpenCode in a fresh worktree) cannot block on interactive prompts when .ai-devkit.json is missing.
  • Renamed debug Skill to dev-debug - The debug skill has been renamed to dev-debug across the registry, templates, and docs to clarify its purpose.
  • Renamed capture-knowledge Skill to document-code - The capture-knowledge skill has been renamed to document-code across the registry, templates, and docs to better reflect its purpose.
  • Built-in Skills Updated - Updated the CLI's built-in skill list to reflect the dev-debug and document-code renames.

Fixed

  • Resumed Claude Session Matching - agent-manager now matches resumed Claude Code sessions by parsing the --resume <uuid> argument, fixing detection of sessions started via claude --resume.

[0.28.0] - 2026-05-09

Added

  • Agent Sessions Command - New ai-devkit agent sessions command to list historical agent sessions (#79).
  • Security Review Skill - Added new security-review skill for security-focused review workflows.
  • Playwright CLI Skill - Added Playwright CLI entry to the skills registry.
  • Google Skills Repository - Added Google skills repository to registry.json.

Changed

  • Strip Claude Code Tag - agent-manager now strips the Claude Code tag from session output.
  • Dev Lifecycle Execute Phase - Improved the dev-lifecycle skill's execute phase guidance.
  • Branding Updates - Updated branding across CLI command templates, command files, and skills.

Fixed

  • AppleScript Escape Utility - Added an AppleScript escape utility in agent-manager to handle quoting safely.
  • Native FS Session Finding - Switched agent-manager session finding to use native fs instead of shelling out via exec.

[0.27.0] - 2026-04-25

Fixed

  • Check Status Path Traversal - Validated FEATURE argument in check-status.sh to prevent path traversal (#66).
  • Shell Injection in TerminalFocusManager - Fixed execAsync calls to use array arguments instead of string interpolation, preventing potential shell injection.

Changed

  • SkillManager Refactor - Split SkillManager into dedicated SkillIndex and SkillRegistry modules for better separation of concerns.
  • ClaudeCodeAdapter Refactor - Extracted session parsing logic from ClaudeCodeAdapter into a standalone ClaudeSessionParser utility.
  • Centralized Error Handling - Added withErrorHandler utility for consistent error handling across CLI commands (agent, channel, memory, skill).
  • AgentManager getAdapter - Added getAdapter() method to AgentManager, reducing adapter resolution duplication in CLI commands.
  • Standardized Error Types - Updated all try-catch blocks to use unknown error type for type safety.
  • Environment Code Consistency - Renamed environment code references for consistency across CLI (init, Config, EnvironmentSelector, TemplateManager).
  • Code Cleanup - Removed stray console.log statements, cleaned up tests, and added .editorconfig.

[0.26.0] - 2026-04-22

Added

  • Gemini CLI Agent Adapter - Added GeminiCliAdapter to the agent manager so ai-devkit agent list and ai-devkit agent detail can discover and inspect running Gemini CLI sessions, with process detection, session discovery via projectHash matching, and content normalization for Gemini's polymorphic message format (#70).
  • Gemini CLI Channel Support - Added GeminiCliAdapter to the channel command for Gemini CLI channel operations.

Fixed

  • Shell Injection Prevention - Hardened git utility functions against shell injection.
  • Skill Path Validation - Added validation for skill paths to prevent invalid path traversal in SkillManager.

Changed

  • Removed Redundant Commands - Cleaned up duplicate command templates (capture-knowledge, debug, simplify-implementation) from both root commands/ and packages/cli/templates/commands/.

[0.25.0] - 2026-04-21

Fixed

  • Agent Manager Terminal.app Detection - Fixed Terminal.app detection in the agent manager.
  • Agent Manager iTerm Sending - Fixed sending input to agents in iTerm.

[0.24.0] - 2026-04-18

Changed

  • Flat Config Structure - Moved registries and skills to top-level fields in project, global, and template configs. Previously registries was nested under skills (skills.registries); now both registries and skills sit at the same level for consistency across all config contexts.
  • Template Registries Support - Init templates can now declare custom registries that get saved to the project config during ai-devkit init --template.
  • Shared Registry Filter - Extracted duplicate registry-filtering logic from ConfigManager and GlobalConfigManager into a shared filterStringRecord() helper.
  • Removed Dead Code - Removed unused getInstalledSkills() method, normalizeSkillsConfig() method, SkillsConfig interface, and SkillRegistriesConfig interface.

Breaking Changes

  • The skills field in .ai-devkit.json is now always a plain array of { registry, name } objects. The previous object format ({ registries: {...}, installed: [...] }) is no longer supported.
  • The global config (~/.ai-devkit/.ai-devkit.json) now uses a top-level registries field instead of skills.registries.

[0.23.1] - 2026-04-17

Fixed

  • Skill Remove Config Cleanup - Fixed skill remove not removing the skill entry from .ai-devkit.json.
  • Install Update Guard - Fixed install service updating config when there are no successful installs.
  • Install Skills Update Guard - Fixed install service including skills in the update object when there are no successful skills.
  • Install Skills Object Format - Fixed install failing when the skills field in config is an object with an installed array rather than a plain array.

[0.23.0] - 2026-04-17

Added

  • MCP Server Config Standardization - Added mcpServers support to init and install flows with environment-specific generators (ClaudeCodeMcpGenerator, CodexMcpGenerator) for standardized MCP server config across AI agents (#61).
  • Channel Debug Flag - Added --debug flag to the channel command for verbose debugging output.

Fixed

  • Memory MCP Tool Names - Renamed MCP tool names from dot-separated (memory.storeKnowledge) to underscore-separated (memory_storeKnowledge, memory_updateKnowledge, memory_searchKnowledge) to comply with strict MCP client naming regex ^[a-zA-Z0-9_-]{1,64}$; backward-compat aliases retained for existing users (#59).
  • Skill Add With No Environments - Fixed skill add crashing when no environments are defined in config.
  • Config addSkill Crash - Fixed addSkill() crashing when the skills field is an object containing a registries key rather than an array.

[0.22.0] - 2026-04-12

Added

  • Channel Connector Package - Added new @ai-devkit/channel-connector package with channel management, config storage, and Telegram adapter support, plus CLI commands for channel operations (#53).
  • Channel Connector Release Workflow - Added publish workflow and package ignore rules for releasing the channel connector package.
  • Codex Plugin Config - Added .codex-plugin/plugin.json configuration.
  • Dev Lifecycle Guide - Added web documentation for the dev-lifecycle skill.
  • Shopify Toolkit Registry Entry - Added the Shopify toolkit skill to the registry.
  • Git Hooks - Added Husky pre-commit and pre-push hooks.

Changed

  • Memory Skill Workflow - Improved the memory skill workflow and updated the OpenAI agent configuration used by the skill.
  • Dev Lifecycle Brainstorming - Updated the dev-lifecycle new-requirement guidance to include brainstorming refinements.
  • Code Review Guidance - Refreshed code review instructions across command templates and dev-lifecycle references.
  • CLI Environment Support - Added Amp Code environment configuration support and updated related CLI expectations.

[0.21.1] - 2026-04-04

Fixed

  • Interactive Skill Add Parsing - Fixed ai-devkit skill add <registry> so it no longer requires skill-name and can open the interactive skill selection flow.

[0.21.0] - 2026-04-04

Added

  • Interactive Skill Selection - ai-devkit skill add can now present an interactive multi-select flow when adding skills from a registry (#51).
  • Memory DB Path Configuration - Added support for configuring a project-specific memory database path and using it in CLI and memory API flows (#50).
  • Senior Engineer Skills - Added verify and tdd skills and included them in the senior-engineer template.
  • Targeted Global Skill Install Prompt - Added environment-aware prompting to support targeted global skill installation.
  • E2E Coverage - Added end-to-end test coverage for the new memory database path configuration flow.

Changed

  • Skill Guidance - Updated memory skill instructions and refreshed skill red-flag guidance.
  • Bundled Skills Data - Added new bundled skills and updated skill registry metadata.
  • E2E Test Maintenance - Cleaned up the end-to-end test suite for the new CLI flows.

Fixed

  • Ora Compatibility - Fixed the ora dependency/version issue.

[0.20.1] - 2026-03-13

Added

  • Agent Orchestration Skill - New agent-orchestration skill for coordinating multi-agent workflows, including OpenAI agent configuration.

Changed

  • TTY Writer - Fixed sending enter separately for more reliable agent input delivery.
  • Dependency Updates - Upgraded project dependencies.
  • Analytics Config - Disabled Nx analytics in project configuration.

[0.20.0] - 2026-03-12

Added

  • Agent Detail Command - New ai-devkit agent detail --id <name> command to inspect running agent conversations (#49).
  • Skill Registry - Added samber/cc-skills-golang skill repository.

Changed

  • Agent Identifier - Updated agent identifier; removed slug field from AgentInfo in favor of simplified name-based matching.

[0.19.0] - 2026-03-11

Added

  • Agent List CWD - Agent list command now displays the current working directory for each running agent (#47).
  • Clarification & Brainstorming Loop - Added clarification and brainstorming loop to review-design and review-requirements commands.

Changed

  • Generalized Session Mapping - Refactored process-to-session mapping into shared utilities (matching, session, process) used by both Claude Code and Codex adapters, replacing adapter-specific implementations (#45).
  • Claude Sessions PID Matching - Updated Claude Code session matching for more reliable PID-based detection (#48).

[0.18.0] - 2026-03-10

Added

  • Custom Docs Directory - Added support for configuring a custom AI documentation directory via paths.docs and ai-devkit init --docs-dir <path>.

Changed

  • Claude Code Adapter - Reimplemented Claude Code session matching to use process start times, bounded session scanning, and direct session-based summaries for more reliable agent detection.
  • Node.js Requirement - Updated the minimum supported Node.js version to 20.20.0.

Fixed

  • Codex Cross-Repo Matching - Prevented Codex agent session matching from incorrectly attaching sessions across repositories.
  • Git Passphrase Prompts - Fixed skill registry git operations so passphrase input works correctly during clone and update flows.

[0.17.0] - 2026-03-09

Added

  • Agent Send Command - New ai-devkit agent send command for sending input to running agents via TTY writer.
  • Agent Type Display - Agent list now shows agent type (e.g., Claude Code, Codex) in the listing output.

Changed

  • Worktrees Location - Updated dev-lifecycle skill worktree setup references.

[0.16.0] - 2026-02-27

Added

  • Codex Adapter - Added Codex adapter support.
  • Agent Manager Package - Added standalone @ai-devkit/agent-manager package.

Changed

  • CLI Agent Migration - Migrated agent command to @ai-devkit/agent-manager.
  • Skill Registry Priority - Updated skill registry priority handling.
  • Init/Install Templates - Stopped copying context templates during init/install.
  • Version Output Source - --version now uses package version output.
  • Documentation - Updated docs and improved web agent setup guide with template-based init.
  • Project Templates - Added and updated templates used across setup flows.
  • Install Validation - Added install smoke-test updates for ai-devkit install.

Fixed

  • Agent List Display - Kept the working-on column to one line in agent list output.
  • Claude PID Session Mapping - Prefer exact history cwd for Claude pid-session mapping.
  • Config Manager Phase Handling - Guarded missing phases in config manager.
  • Docs Typos/Troubleshooting - Fixed typo and added Codex sandbox npx troubleshooting FAQ.

[0.15.0] - 2026-02-24

Added

  • Install Command - Added ai-devkit install to apply project configuration from .ai-devkit.json
    • Supports --config <path> for custom config file locations
    • Supports --overwrite for non-interactive full overwrite mode
    • Installs environments, phases, and skills in a single run with summary output

[0.14.0] - 2026-02-21

Changed

  • Dependency Updates - Upgraded better-sqlite3

[0.13.0] - 2026-02-20

Added

  • Lint Command - Added ai-devkit lint command support
  • Template Mode for Init - Added init template mode with YAML/JSON support
  • Memory Update Command - Added ai-devkit memory update for modifying knowledge items by ID
  • New Skills - Added capture-knowledge, simplify-implementation and technical-writer skills
  • Plugin Support - Added .claude-plugin and .cursor-plugin integration files

Changed

  • Dev Lifecycle Workflows - Refactored worktree setup and new-requirement flow
  • Lifecycle Documentation - Updated docs to require feature worktrees and make bootstrap language-agnostic
  • Web Docs Navigation - Added linkable anchors for documentation section headings
  • Command Templates - Updated CLI command templates
  • Skill Registry Handling - Refresh cached skill registry automatically on skill add
  • Documentation Updates - Refreshed README and development docs for current CLI behavior

Fixed

  • Memory Test Stability - Fixed flaky updated_at timestamp test in memory module

[0.12.0] - 2026-02-17

Added

  • Dev Lifecycle Skill - Added structured SDLC skill with phase references and helper scripts
  • Debug Skill - Added reusable debug skill definitions for agent workflows
  • Web Skill Search Experience - Added /skills web page and related docs/navigation updates
  • Memory Search Table Output - Added ai-devkit memory search --table for terminal-friendly results

Changed

  • Skill Registry Data - Updated skill registry/index content and automated rebuild outputs
  • Documentation - Added/updated AI phase docs for setup wizard, web skill search, and memory search table output

Fixed

  • Init Environment Parsing - Improved init -e handling for full environment values

[0.11.0] - 2026-02-06

Added

  • Skill Search - New skill find command to discover skills across all registries
    • Keyword Search: Find skills by name or description (e.g., ai-devkit skill find typescript)
  • Skill Index Rebuild - New skill rebuild-index command for search feature

Changed

  • Native Fetch - Migrated network calls from https to native fetch API for cleaner code
  • GITHUB_TOKEN Support - GitHub API calls now use GITHUB_TOKEN environment variable when available

[0.10.0] - 2026-02-01

Added

  • Agent Management - Detect and control external AI agents
    • List Agents: ai-devkit agent list - View running agents (Claude Code, etc.)
    • Open Agent: ai-devkit agent open <name> - Focus agent terminal window
    • Terminal Support: Works with tmux, iTerm2, and Apple Terminal
    • Fuzzy Matching: Open agents by partial name

[0.9.0] - 2026-01-28

Added

  • Terminal UI Standardization - Centralized terminal output utility for consistent CLI experience
  • Skill Update Command - New ai-devkit skill update command for updating skills from registries
    • Update All Skills: ai-devkit skill update - Updates all cached skill registries via git pull
    • Update Specific Registry: ai-devkit skill update <registry-id> - Updates only the specified registry (e.g., ai-devkit skill update anthropic/skills)

Changed

  • Module Resolution - Updated TypeScript configuration from Node16 to CommonJS for better compatibility

Fixed

  • Graceful Exit - Commands now properly exit with code 0 on successful completion
    • skill list - Added explicit process.exit(0) when no skills found
    • skill remove - Added explicit process.exit(0) after successful removal

[0.8.1] - 2026-01-26

Added

  • Custom Skill Registries - Support skills.registries in global ~/.ai-devkit/.ai-devkit.json for adding multiple registries that merge with defaults and override on conflicts.
  • Global Registry Reader - New global config reader for resolving custom registries in skill commands.

Changed

  • Skill Registry Resolution - Skill commands now merge default and custom registries, with offline cache fallback when a registry URL is not configured.

[0.8.0] - 2026-01-26

Added

  • Memory Skill Template - New skill for integrating memory service capabilities into agent workflows
  • Comprehensive Documentation - Added extensive documentation pages for:
    • Getting Started guide
    • Supported AI agents reference
    • Development with AI DevKit
    • Debug workflows
    • Understanding existing code
    • Memory service usage
    • Skills management
  • Updated base template for all environments

[0.7.0] - 2026-01-25

Added

  • Skill Management - Centralized registry for managing Agent Skills across projects
    • One-Command Installation: ai-devkit skill add <registry>/<repo> <skill-name>
    • Local Cache: Skills stored in ~/.ai-devkit/skills/ to avoid duplication
    • Symlink-First Strategy: Symlinks with automatic copy fallback for Windows
    • Multi-Environment Support: Works with Cursor, Claude Code, Codex, OpenCode, and Antigravity
    • CLI Commands:
      • ai-devkit skill add <registry>/<repo> <skill-name> - Install a skill from registry
      • ai-devkit skill list - List all installed skills with sources
      • ai-devkit skill remove <skill-name> - Remove skill from project
    • Features:
      • Centralized registry file (skills/registry.json) with verified repositories
      • Automatic .ai-devkit.json creation if missing
      • Environment filtering (only shows/uses environments with skill support)
      • Git repository caching for efficient reuse across projects
      • Validation for registry IDs and skill names (follows Agent Skills spec)

[0.6.0] - 2026-01-22

Added

  • Knowledge Memory Service (packages/memory) - A lightweight MCP-based memory service for AI agents
    • Store and retrieve actionable knowledge using SQLite with FTS5 full-text search
    • Core Features:
      • 🔍 Full-Text Search - FTS5 with BM25 ranking
      • 🏷️ Tag-Based Filtering - Boost results by contextTags
      • 📁 Scoped Knowledge - global, project, or repo-specific rules
      • 🔄 Deduplication - Prevents duplicate content
    • CLI Integration: New memory command family
      • ai-devkit memory store - Store new knowledge items
      • ai-devkit memory search - Search for relevant knowledge
  • Global Setup Command - New ai-devkit setup --global command for installing commands globally
    • Copy AI DevKit commands to global environment folders
    • Support for Antigravity (~/.gemini/antigravity/global_workflows/) and Codex (~/.codex/prompts/)
    • Interactive environment selection with only global-capable environments shown
    • Overwrite prompts for existing global commands
    • Cross-platform support using os.homedir() and path.join()

[0.5.0] - 2025-01-15

Added

  • Antigravity Support - Added support for Google Antigravity
  • New Slash Command - /simplify-implementation for analyzing and simplifying existing implementations

Changed

  • Dynamic TOML Generation - Refactored TemplateManager to dynamically generate .toml files from .md files at runtime

[0.4.2] - 2025-11-05

  • Fixed Gemini CLI integration #3
  • Added test for TemplateManager.ts
  • Fixed Github Copilot integration #4

[0.4.0] - 2025-10-31

Added

  • Multi-Environment Setup - Support for 10 AI development environments
    • Interactive environment selection with multi-choice prompts
    • Support for Cursor, Claude Code, GitHub Copilot, Google Gemini, OpenAI Codex, Windsurf, KiloCode, AMP, OpenCode, and Roo Code
    • Unified template structure with AGENTS.md files for all environments
    • Environment-specific command directories and configuration files
    • Override protection with confirmation prompts for existing environments
    • Config persistence storing selected environments array

Changed

  • Breaking Changes - Removed legacy single-environment support for cleaner API
    • Renamed EnvironmentId to EnvironmentCode throughout codebase
    • Removed legacy Environment type union (cursor | claude | both)
    • Updated config schema to use environments: EnvironmentCode[]
    • All environments now use standardized AGENTS.md context files

Technical Improvements

  • Testing Infrastructure - Complete test suite implementation
  • Architecture - Modular design improvements

[0.3.0] - 2025-10-15

Added

  • /debug - Structured assistant for clarifying issues, analyzing options, and agreeing on a fix plan before coding
  • /capture-knowledge - Analyze and explain how code works from any entry point
    • Supports file, folder, function, and API endpoint analysis
    • Recursive dependency analysis with configurable depth (max: 3)
    • Automatic generation of mermaid diagrams (flowcharts, sequence, architecture, class diagrams)
    • Knowledge capture documentation saved to docs/ai/implementation/knowledge-{feature-name}.md
    • Visual dependency tree and component relationship mapping
    • Includes error handling, performance considerations, and improvement suggestions

[0.2.0] - 2025-10-14

Added

  • Eight slash commands for Cursor and Claude Code:
    • /new-requirement - Complete guided workflow from requirements to PR/MR creation
    • /code-review - Structured local code reviews
    • /execute-plan - Walk feature plans task-by-task
    • /writing-test - Generate tests with guidance for 100% coverage
    • /update-planning - Reconcile progress with planning docs
    • /check-implementation - Compare implementation with design
    • /review-design - Review system design and architecture
    • /review-requirements - Review and summarize requirements
  • Claude workspace configuration file (CLAUDE.md)
  • Cursor rules file (ai-devkit.md)
  • Design documentation requirements for mermaid diagrams (architecture and data flow)

[0.1.0] - 2025-10-14

Added

  • Initial release of AI DevKit CLI
  • Interactive init command for project initialization
  • Support for Cursor and Claude Code environments
  • Seven phase templates: requirements, design, planning, implementation, testing, deployment, monitoring
  • phase command for adding individual phases
  • Configuration management with .ai-devkit.json
  • Template overwrite prompts for existing files
  • Comprehensive documentation and README
  • TypeScript support with full type definitions
  • Cursor rules in .cursor/rules/ directory
  • Cursor slash commands as individual Markdown files in .cursor/commands/
  • Claude Code workspace configuration in CLAUDE.md

Features

  • Interactive prompts with Inquirer
  • Flag-based overrides for automation
  • Markdown templates with YAML frontmatter
  • Cursor rules and slash commands generation
  • Claude Code workspace configuration
  • State tracking for initialized phases