|
| 1 | +# Changelog |
| 2 | + |
| 3 | +All notable changes to PlanDB are documented here. |
| 4 | + |
| 5 | +Format: [Keep a Changelog](https://keepachangelog.com/en/1.1.0/). Versioning: [SemVer](https://semver.org/). |
| 6 | + |
| 7 | +--- |
| 8 | + |
| 9 | +## [0.2.0] — 2026-03-26 |
| 10 | + |
| 11 | +The "issue tracker for AI agents" release. Four platform features, production CI/CD, and a complete repositioning. |
| 12 | + |
| 13 | +### Added |
| 14 | +- **Context store with BM25 search** — `plandb context "..." --kind discovery` records project knowledge. `plandb search "query"` does BM25-ranked search across context entries and task descriptions. Freeform `--kind` labels. Available on CLI, MCP, and HTTP. |
| 15 | +- **Lazy recall** — `plandb go` auto-surfaces relevant context entries for the claimed task. Agents get knowledge without explicitly searching. |
| 16 | +- **Task lifecycle hooks** — `--pre-hook` and `--post-hook` shell commands on tasks. Fire at state transitions with env vars (`PLANDB_TASK_ID`, `PLANDB_TASK_TITLE`, etc.). Advisory — failures warn, never block. |
| 17 | +- **Contextual action hints** — every CLI response teaches the agent what to do next. Kind-aware: research tasks hint "record discoveries", code tasks hint "split if complex". |
| 18 | +- **Agent-scale diagram** (`assets/agent-scale.png`) — panoramic visualization of decomposition, parallelization, and mid-flight adaptation. |
| 19 | +- **CI workflow** (`ci.yml`) — runs on every push/PR: build, test, functional tests, clippy, fmt. |
| 20 | +- **Cargo caching** in CI for faster builds. |
| 21 | + |
| 22 | +### Changed |
| 23 | +- **CLI prompt redesigned** — 51% smaller (199 → 110 lines, ~2400 → ~1200 tokens). Example-first instead of documentation-heavy. Agents now create proper dependency graphs (validated with Codex and Gemini). |
| 24 | +- **Release workflow** — releases now trigger on version tags (`v*`) only, not every push to main. Release names use "PlanDB" (was "Plandb"). |
| 25 | +- **README repositioned** — "The issue tracker your AI agents are missing. Think Linear or Jira — but for your agents." Mapping table, thought leadership section, showcase moved up. |
| 26 | +- **All prompts synced** — CLI, MCP, HTTP, and `--help` all reference context store, lazy recall, and hooks. |
| 27 | +- **Branch protection** enabled on main: require CI pass, no force push, no branch deletion. |
| 28 | + |
| 29 | +### Fixed |
| 30 | +- `plandb add --dep t-nonexistent` no longer creates orphaned tasks. Validates all dependency targets upfront with fuzzy-match suggestions. |
| 31 | +- `batch_create_tasks` now correctly passes `pre_hook`/`post_hook` parameters (was passing 33 params, needed 35). |
| 32 | +- Context `--kind` default corrected to "discovery" (was "note"). |
| 33 | +- Help text references fixed from `--type` to `--kind`. |
| 34 | +- Removed references to `plandb resume` and `plandb templates` from prompts (commands don't exist yet). |
| 35 | + |
| 36 | +### Removed |
| 37 | +- 10 stale branches cleaned up. |
| 38 | +- Template references removed from all agent-facing prompts (feature not yet shipped). |
| 39 | + |
| 40 | +--- |
| 41 | + |
| 42 | +## [0.1.x] — 2026-03-04 to 2026-03-25 |
| 43 | + |
| 44 | +Foundation releases. Core task graph, compound graph (containment + dependencies), recursive decomposition, mid-flight adaptation primitives, multi-agent coordination, quality gates, CLI/MCP/HTTP interfaces, cross-platform binary builds. |
| 45 | + |
| 46 | +See [GitHub releases](https://github.com/Agent-Field/plandb/releases) for per-patch details. |
0 commit comments