# Changelog ## 3.0.0 ### Major Changes - Jensen 3.0.0 establishes the durable distributed autonomous agent runtime foundation. It adds durable Missions with restart-safe lifecycle and completion authority, Scheduler/Assignment/Worker execution, remote execution and capability routing, modern MCP support, shared-inference admission for local and remote agents, logical subagents, durable orchestration DAGs, and Governance/Budgets with resource-pressure and Jensen-owned process stewardship. This release changes the public runtime boundary: execution state is authoritative in Jensen-owned durable stores, model completion is only a candidate until deterministic verification and the Completion Gate accept it, and distributed execution requires explicit mission, assignment, worker, and remote-target configuration. Existing 2.x sessions and provider/model configuration remain readable where their persisted schema is supported; operators should review remote-execution, shared-inference, and governance settings before enabling those capabilities. Invalid or incompatible durable state fails closed rather than being silently converted to success. Jensen 3.0.0 does not include the future persistent daemon/control plane, mobile or trusted remote clients, voice/STT/TTS, device capabilities, or ambient-assistant services. Those remain planned for later Jensen 3.x phases. ### Patch Changes - Updated dependencies [6fd00ec] - Updated dependencies [296374a] - Updated dependencies [baed742] - Updated dependencies - Updated dependencies [4bc8baf] - Updated dependencies [c58ac4b] - Updated dependencies [be46f86] - @apholdings/jensen-code@3.0.0 - @apholdings/jensen-agent-core@3.0.0 - @apholdings/jensen-ai@3.0.0 ## 2.1.0 ### Minor Changes - 392db26: Jensen 2.1.0 — Reliable Agent Runtime. Introduces the Reliability Kernel, which moves execution authority from the model into Jensen. The model proposes actions and completion; Jensen owns state, execution, evidence, validation, and completion. - Mission Ledger: durable, machine-readable, session-associated mission state (goal, constraints, acceptance criteria, evidence, blockers) that survives process exit, `jensen resume `, and context compaction. - Structured agent actions: model output is normalized into a finite, Jensen-owned action union (`tool_call`, `request_context`, `mission_update`, `final_candidate`, `blocked`, `no_op`). - Action validation: every executable action is validated before execution (tool exists, schema, required arguments, boundary, permission, forbidden actions); invalid actions are rejected with structured failures and never run. - Evidence Store: records what Jensen actually observed, never model claims; deterministic verification evidence is authoritative. - Verification Engine: deterministic build/test/file/search/git-diff verifiers. - Acceptance criteria: user-required and system-derived criteria with evidence-gated status; a model assertion can never mark a criterion passed. - Completion Gate: `FINAL_CANDIDATE` is a proposal, not completion; Jensen rejects premature completion with `FINALIZATION_REJECTED`. - Reliability Suite: deterministic adversarial scenarios (R01–R15) prove the trusted runtime stays safe when the model misbehaves. 2.1.0 introduces the Reliability Kernel; advanced autonomous retry/recovery policy remains future work (2.1.1). ### Patch Changes - Updated dependencies [392db26] - @apholdings/jensen-agent-core@2.1.0 - @apholdings/jensen-ai@2.1.0 - @apholdings/jensen-code@2.1.0 ## 2.0.2 ### Patch Changes - 10549ec: Add explicit session resume by ID. `jensen resume ` resolves a persisted session by exact ID, restores its conversation history, model/thinking configuration, memory/todos/tasks, and working directory, and continues writing under the same session ID rather than creating a new one. - Exact session resolution (never prefix matching, never fall back to the latest session). - Same-session continuation across repeated resume cycles. - Safe failure for unknown IDs (`Session not found: `), corrupted session files (persisted bytes are never overwritten), and path-like input (never treated as a filesystem path). - Backward compatible with sessions persisted by 2.0.1 and earlier versions. - New regression coverage for ID lookup, file validation, continuation, compaction continuity, and end-to-end hydration through `createAgentSession`. - Updated dependencies [10549ec] - @apholdings/jensen-ai@2.0.2 - @apholdings/jensen-agent-core@2.0.2 - @apholdings/jensen-code@2.0.2 ## 2.0.1 ### Patch Changes - 1854dad: Refresh the OpenRouter model catalog from the live OpenRouter Models API. Regenerates `models.generated.ts` against `https://openrouter.ai/api/v1/models` so every currently tool-capable OpenRouter model resolves through Jensen's existing OpenAI-compatible OpenRouter provider. Highlights from the refresh: - Added support for newly available OpenRouter models such as `x-ai/grok-4.6`, `qwen/qwen3.8-max`, `thinkingmachines/inkling-small`, `sakana/sakana-namazu`, `upstage/solar-pro4`, and `~deepseek/deepseek-v4-flash-latest`. - Removed OpenRouter entries that the live API no longer returns (`openai/gpt-5.1-chat`, `openai/gpt-5.3-chat`, `inclusionai/ling-3.0-flash:free`). - Dropped stale manual pricing overrides for `moonshotai/kimi-k2.5` and `z-ai/glm-5`; their metadata now comes straight from the live API. The OpenRouter provider architecture is unchanged: `openrouter` models continue to route through the existing `openai-completions` provider with the same reasoning-effort, provider-routing, and API-key handling as before. - Updated dependencies [1854dad] - @apholdings/jensen-ai@2.0.1 - @apholdings/jensen-agent-core@2.0.1 - @apholdings/jensen-code@2.0.1 ## 2.0.0 ### Minor Changes - 1f51e24: Add durable Mission Graph and multi-repository program execution. Introduces a durable, versioned, scope-explicit execution graph in `@apholdings/jensen-code` (`jensen mission …`, `jensen doctor mission`) for orchestrating work across one or more declared repositories: - A versioned, hashed **Mission Graph** whose identity is a canonical sha-256 of its semantic payload; revisions are monotonic, completed history is preserved, and a mission can never expand its own scope. - Authoritative **repository identity** (canonical, credential-free) and **worktree allocation** with isolation boundaries; symlink / junction escape and any allocation of the operator worktree are rejected; undeclared repository drift is detected. - A dependency-aware **scheduler** with deterministic parallel waves, a configurable parallelism bound, per-repository **write-conflict serialization**, **critical-path** analysis, and mission/objective budget bounds. - Typed **contracts** linking producers to consumers with compatibility and stale-contract detection, and transactional **integration** (checkpoint → validate → confirm) with **localized rollback** that preserves independent completed work. - **Human approval gates** (no self-approval, principal-verified, scope enforced, rejectable, expirable) and **external blockers** that require concrete evidence and cannot be fabricated. - Objective and mission **state machines**, repository-scoped **leases**, atomic durable storage, an append-only replayable **event log**, and **reboot recovery**: a reboot never duplicates an objective, commit, PR or publication, treats a recorded dead process as `missing` (never reusing stale process or lease authority), and preserves independent completed work. Safety invariants preserved: objectives cannot grant one another authority, approval nodes cannot be auto-approved, external blockers cannot be fabricated, routing never overrides mission policy, evaluation cannot grant authority, and replay has zero external effects. All CLI/diagnostic paths run offline on deterministic fixtures in normal CI. ### Patch Changes - Updated dependencies [1f51e24] - @apholdings/jensen-ai@2.0.0 - @apholdings/jensen-agent-core@2.0.0 - @apholdings/jensen-code@2.0.0 ## 1.9.0 ### Minor Changes - 0432891: Add evidence-driven adaptive orchestration subsystem. Introduces a canonical, evidence-driven orchestration decision engine in `@apholdings/jensen-code` (`jensen routing …`) that selects the execution strategy for each task using durable, replayable evidence rather than hardcoded workflows: - Deterministic task feature extraction with a versioned schema. - Candidate generation from canonical registries only, bounded and deterministically ordered, with hard policy filtering (workspace, tool effects, network, provider/model allowlists, cost, model-call, subagent, local-only, live-provider authorization). - A deterministic baseline policy with explicit rule IDs and precedence that works offline and during evaluation-store failure. - Evaluation-informed scoring with explicit uncertainty and missing-evidence handling (missing evidence is never zero); safety failures stay separate hard constraints and are never averaged away. - Multi-objective selection across correctness/safety/reliability/cost/latency with explicit, versioned weights and operator-authoritative objective. - Zero-effect shadow routing and labeled counterfactual evaluation. - Conservative offline policy optimization from content-addressed datasets with explicit safety/correctness/flakiness promotion gates and idempotent rollback. - Typed escalation / de-escalation and fallback / degradation with hard bounds, operator budget ceilings, and a mandatory finalization reserve. - Budget-class selection, skill/subagent topology selection, retrieval strategy selection, and long-horizon integration at governed phase boundaries. - Durable addressable events, replayable decisions, and deterministic bounded drift detection (quality/cost/latency/failure-cluster/retrieval/flakiness/ policy-selection). - CLI (`jensen routing …`, `jensen doctor routing`), versioned RPC operations, and dashboard projections. Authority hierarchy: user authorization > safety/effect policy > workspace and transaction authority > hard constraints > operator overrides > validated policy > evaluation-backed scores > heuristics > model recommendation. Safety invariants preserved: routing never overrides safety, never grants new tool authority, never expands workspace scope, never exceeds the operator budget, never enables a live provider implicitly, and never promotes a policy automatically. All CLI/RPC/diagnostic paths run offline on deterministic fixtures in normal CI (no paid API calls). ### Patch Changes - Updated dependencies [0432891] - @apholdings/jensen-ai@1.9.0 - @apholdings/jensen-agent-core@1.9.0 - @apholdings/jensen-code@1.9.0 ## 1.8.3 ### Patch Changes - Updated dependencies [b0fb9ee] - @apholdings/jensen-code@1.8.3 - @apholdings/jensen-agent-core@1.8.3 - @apholdings/jensen-ai@1.8.3 ## 1.8.2 ### Patch Changes - 3f46ae6: Complete the production evaluation runtime with real sandbox lifecycle enforcement, opt-in live-provider execution, independent reviewers, paired Cavecrew comparisons, versioned evaluation RPC operations, bounded dashboard projections, protected retention and pruning, and single-commit release provenance checks. - Updated dependencies [3f46ae6] - @apholdings/jensen-agent-core@1.8.2 - @apholdings/jensen-ai@1.8.2 - @apholdings/jensen-code@1.8.2 ## 1.8.1 ### Patch Changes - Harden the evaluation artifact store and doctor exit semantics, complete deterministic replay, stability, retrieval, clustering, pruning, RPC, and dashboard projections, and make binary builds resolve Playwright's Chromium BiDi dependency from a clean checkout with verified release manifests. - Updated dependencies - @apholdings/jensen-agent-core@1.8.1 - @apholdings/jensen-ai@1.8.1 - @apholdings/jensen-code@1.8.1 ## 1.8.0 ### Minor Changes - 7447be7: Add a versioned evaluation runtime with deterministic scenario packs, isolated fixtures, replay-safe artifacts, baseline comparison, safety release gates, metrics, and the `jensen eval` CLI. ### Patch Changes - Updated dependencies [7447be7] - @apholdings/jensen-ai@1.8.0 - @apholdings/jensen-agent-core@1.8.0 - @apholdings/jensen-code@1.8.0 ## 1.7.1 ### Patch Changes - Updated dependencies [138f6a8] - @apholdings/jensen-code@1.7.1 - @apholdings/jensen-agent-core@1.7.1 - @apholdings/jensen-ai@1.7.1 ## 1.7.0 ### Minor Changes - 6a9e948: Semantic workspace intelligence (1.7.0) Added durable, deterministic workspace indexing and hybrid lexical/symbolic/ semantic retrieval: - Workspace identity resolution (symlink/junction-safe, Git-repository and worktree aware, secret-free). - Durable per-workspace SQLite index on the built-in `node:sqlite` backend with atomic, versioned, rebuildable generations. No new dependency; no cloud or paid service required. - Deterministic file discovery and classification with `.gitignore` / `.jensenindexignore` support, binary/vendor/cache exclusion, and conservative secret-sensitive-file exclusion (never chunked, embedded, or returned). - Content-addressed, syntax-aware chunking (symbol/section/config/fallback). - Portable postings + BM25 lexical retrieval with identifier-aware tokenization (exact/prefix/phrase/path/symbol) that does not depend on FTS5 availability. - Symbolic index and relationships with heuristic parser plus LSP integration. - Bounded Git metadata and worktree fingerprints for ranking/invalidation. - Pluggable embedding backends: deterministic local fixture (default, no paid dependency), local loopback OpenAI-compatible endpoint, and opt-in remote. Remote embedding is disabled by default. Deterministic exact vector search. - Deterministic query planner, reciprocal-rank fusion, and heuristic reranking with explicit reason codes and per-file diversity. - Freshness labeling, content-hash revalidation, and incremental refresh (added/modified/removed files), plus pruning and integrity verify/rebuild. - CLI commands (`index`, `search`, `retrieval`, `doctor index|embeddings|retrieval`), provider-independent workspace tools with declared effects, doctor checks, bounded context packets and evidence records, and per-subagent retrieval policies (scout, cavecrew-investigator, planner, worker, builder, reviewer). - Deterministic Linux and Windows test suites; requires Node.js >= 22.5.0 (bundled `node:sqlite`). The index is always a disposable, rebuildable projection and never an execution authority. ### Patch Changes - Updated dependencies [6a9e948] - @apholdings/jensen-code@1.7.0 - @apholdings/jensen-agent-core@1.7.0 - @apholdings/jensen-ai@1.7.0 ## 1.6.2 ### Patch Changes - b565a26: Add a canonical policy-bound subagent registry with explicit OpenRouter model routing, typed unknown-agent resolution, structured output contracts, and validated Cavecrew investigator, builder, and reviewer roles. - 708dac4: Wire canonical subagent resolution, isolated context packets, parent output validation, and deterministic transactional Cavecrew orchestration into runtime dispatch. Make release artifact integrity checks build their inputs before verification. - Updated dependencies [b565a26] - Updated dependencies [708dac4] - @apholdings/jensen-agent-core@1.6.2 - @apholdings/jensen-ai@1.6.2 - @apholdings/jensen-code@1.6.2 ## 1.6.1 ### Patch Changes - fb0f065: Complete operability replay, safe re-execution, MCP transports, validation, storage diagnostics, and release-state classification for the 1.6 contract. - Updated dependencies [fb0f065] - @apholdings/jensen-agent-core@1.6.1 - @apholdings/jensen-ai@1.6.1 - @apholdings/jensen-code@1.6.1 ## 1.6.0 ### Minor Changes - 24c2d95: Add canonical run observability projections, deterministic render and projection replay, evidence and diagnostic inspection, sanitized support-bundle previews, and conservative MCP configuration/schema/effect validation surfaces. ### Patch Changes - Updated dependencies [24c2d95] - @apholdings/jensen-agent-core@1.6.0 - @apholdings/jensen-ai@1.6.0 - @apholdings/jensen-code@1.6.0 ## 1.5.0 ### Patch Changes - Updated dependencies [af149b0] - @apholdings/jensen-code@1.5.0 - @apholdings/jensen-agent-core@1.5.0 - @apholdings/jensen-ai@1.5.0 ## 1.4.0 ### Patch Changes - Updated dependencies [3485c3c] - @apholdings/jensen-code@1.4.0 - @apholdings/jensen-agent-core@1.4.0 - @apholdings/jensen-ai@1.4.0 ## 1.3.0 ### Patch Changes - Updated dependencies [a65bbd5] - @apholdings/jensen-agent-core@1.3.0 - @apholdings/jensen-code@1.3.0 - @apholdings/jensen-ai@1.3.0 ## 1.2.1 ### Patch Changes - Updated dependencies [cbd624d] - @apholdings/jensen-code@1.2.1 - @apholdings/jensen-agent-core@1.2.1 - @apholdings/jensen-ai@1.2.1 ## 1.2.0 ### Patch Changes - Updated dependencies [328c049] - @apholdings/jensen-code@1.2.0 - @apholdings/jensen-agent-core@1.2.0 - @apholdings/jensen-ai@1.2.0 ## 1.1.12 ### Patch Changes - Updated dependencies [64aab5a] - @apholdings/jensen-ai@1.1.12 - @apholdings/jensen-agent-core@1.1.12 - @apholdings/jensen-code@1.1.12 ## 1.1.11 ### Patch Changes - dcc565a: Prevent Windows shell commands from hanging after launching persistent background processes, while preserving output, timeout, and process-cleanup semantics. - Updated dependencies [dcc565a] - @apholdings/jensen-agent-core@1.1.11 - @apholdings/jensen-ai@1.1.11 - @apholdings/jensen-code@1.1.11 ## 1.1.10 ### Patch Changes - Updated dependencies [1dfcae4] - Updated dependencies [4caca50] - Updated dependencies [9b61ea2] - Updated dependencies [c2bfd7f] - Updated dependencies [f2bfa57] - Updated dependencies [c9be929] - @apholdings/jensen-agent-core@1.1.10 - @apholdings/jensen-ai@1.1.10 - @apholdings/jensen-code@1.1.10 ## 1.1.9 ### Patch Changes - Updated dependencies [1419cfb] - @apholdings/jensen-code@1.1.9 - @apholdings/jensen-agent-core@1.1.9 - @apholdings/jensen-ai@1.1.9 ## 1.1.8 ### Patch Changes - Updated dependencies [2c2ceff] - @apholdings/jensen-code@1.1.8 - @apholdings/jensen-agent-core@1.1.8 - @apholdings/jensen-ai@1.1.8 ## 1.1.7 ### Patch Changes - c322b08: Fix packaging: remove broken `main` and `types` entrypoints that referenced non-existent `dist/index.js` and `dist/index.d.ts`. Mom is a CLI-only package (bin: `mom`). - Updated dependencies [0a60417] - @apholdings/jensen-code@1.1.7 - @apholdings/jensen-agent-core@1.1.7 - @apholdings/jensen-ai@1.1.7 ## [Unreleased] ## 1.1.6 ### Patch Changes - Updated dependencies - @apholdings/jensen-ai@1.1.6 - @apholdings/jensen-agent-core@1.1.6 - @apholdings/jensen-code@1.1.6 ## 1.1.5 ### Patch Changes - Updated dependencies - @apholdings/jensen-code@1.1.5 - @apholdings/jensen-agent-core@1.1.5 - @apholdings/jensen-ai@1.1.5 ## 1.1.4 ### Patch Changes - Updated dependencies - @apholdings/jensen-code@1.1.4 - @apholdings/jensen-agent-core@1.1.4 - @apholdings/jensen-ai@1.1.4 ## 1.1.2 ### Patch Changes - Updated dependencies - @apholdings/jensen-code@1.1.2 - @apholdings/jensen-agent-core@1.1.2 - @apholdings/jensen-ai@1.1.2 ## 1.1.1 ### Patch Changes - Updated dependencies - @apholdings/jensen-code@1.1.1 - @apholdings/jensen-agent-core@1.1.1 - @apholdings/jensen-ai@1.1.1 ## 1.1.0 ### Patch Changes - Updated dependencies - @apholdings/jensen-code@1.1.0 - @apholdings/jensen-agent-core@1.1.0 - @apholdings/jensen-ai@1.1.0 ## 1.0.9 ### Patch Changes - Updated dependencies - @apholdings/jensen-code@1.0.9 - @apholdings/jensen-agent-core@1.0.9 - @apholdings/jensen-ai@1.0.9 ## [1.1.3] - 2026-07-09 ## [1.0.8] - 2026-06-26 ### Patch Changes - Updated dependencies - @apholdings/jensen-ai@1.0.8 - @apholdings/jensen-agent-core@1.0.8 - @apholdings/jensen-code@1.0.8 ## 1.0.7 ### Patch Changes - Updated dependencies - @apholdings/jensen-ai@1.0.7 - @apholdings/jensen-agent-core@1.0.7 - @apholdings/jensen-code@1.0.7 ## 1.0.5 ### Patch Changes - Removed orphaned `WorkingContextPanel` TUI component and its test file, completing the working-context panel removal cleanup. No breaking API changes. - Updated dependencies - @apholdings/jensen-agent-core@1.0.5 - @apholdings/jensen-code@1.0.5 - @apholdings/jensen-ai@1.0.5 ## 1.0.4 ### Patch Changes - Updated dependencies - @apholdings/jensen-agent-core@1.0.4 - @apholdings/jensen-code@1.0.4 - @apholdings/jensen-ai@1.0.4 ## 1.0.3 ### Patch Changes - Release: Add powershell tool, memory snapshots, working-context surface, get_working_context RPC, and /ultraplan command - Updated dependencies - @apholdings/jensen-code@1.0.3 - @apholdings/jensen-agent-core@1.0.3 - @apholdings/jensen-ai@1.0.3 ## 1.0.2 ### Patch Changes - Updated dependencies [becdcc3] - @apholdings/jensen-code@1.0.2 - @apholdings/jensen-agent-core@1.0.2 - @apholdings/jensen-ai@1.0.2 ## 0.57.7 ### Patch Changes - Refactored interactive mode components to use a new BorderedBox for a cleaner, transparent UI with rounded borders in tool executions and user messages. - Updated dependencies - @apholdings/jensen-agent-core@0.57.7 - @apholdings/jensen-ai@0.57.7 - @apholdings/jensen-code@0.57.7 ## [1.1.3] - 2026-07-09 ## [0.57.6] - 2026-03-26 ## 0.57.4 ### Patch Changes - Updated dependencies - @apholdings/jensen-code@0.57.4 - @apholdings/jensen-agent-core@0.57.4 - @apholdings/jensen-ai@0.57.4 ## 0.57.3 ### Patch Changes - Updated dependencies - @apholdings/jensen-code@0.57.3 - @apholdings/jensen-agent-core@0.57.3 - @apholdings/jensen-ai@0.57.3 ## 0.57.2 ### Patch Changes - Updated dependencies - @apholdings/jensen-code@0.57.2 - @apholdings/jensen-agent-core@0.57.2 - @apholdings/jensen-ai@0.57.2 ## 0.1.3 ### Patch Changes - Promote `JENSEN.md` as the preferred project instruction file, keep `AGENTS.md` as a supported fallback with deprecation diagnostics, and update the related interactive messaging and documentation. - Updated dependencies - @apholdings/jensen-ai@0.1.3 - @apholdings/jensen-agent-core@0.1.3 - @apholdings/jensen-code@0.1.3 ## 0.1.0 ### Minor Changes - 2f3d37c: chore: version bump all packages ### Patch Changes - a55721e: Normalize package publish metadata and internal dependency ranges for the Changesets-based release flow. - Updated dependencies [a55721e] - Updated dependencies [2f3d37c] - @apholdings/jensen-ai@0.1.0 - @apholdings/jensen-code@0.1.0 - @apholdings/jensen-agent-core@0.1.0 ## [0.1.5] - 2026-03-21 ## [0.57.5] - 2026-03-24 ## [0.0.1] - 2026-03-15 - Reset version for Jensen Code fork - Updated branding to Jensen Code - Synchronized with monorepo baseline --- ## Inherited Upstream History (pi-mono) ## [0.57.1] - 2026-03-07 ## [0.57.0] - 2026-03-07 ## [0.56.3] - 2026-03-06 ## [0.56.2] - 2026-03-05 ## [0.56.1] - 2026-03-05 ## [0.56.0] - 2026-03-04 ## [0.55.4] - 2026-03-02 ### Fixed - Fixed mom startup crash caused by settings API drift by using `SettingsManager` with workspace-backed storage ([#1444](https://github.com/badlogic/pi-mono/issues/1444)) ## [0.55.3] - 2026-02-27 ## [0.55.2] - 2026-02-27 ## [0.55.1] - 2026-02-26 ## [0.55.0] - 2026-02-24 ## [0.54.2] - 2026-02-23 ## [0.54.1] - 2026-02-22 ## [0.54.0] - 2026-02-19 ## [0.53.1] - 2026-02-19 ## [0.53.0] - 2026-02-17 ## [0.52.12] - 2026-02-13 ## [0.52.11] - 2026-02-13 ## [0.52.10] - 2026-02-12 ## [0.52.9] - 2026-02-08 ## [0.52.8] - 2026-02-07 ## [0.52.7] - 2026-02-06 ## [0.52.6] - 2026-02-05 ## [0.52.5] - 2026-02-05 ## [0.52.4] - 2026-02-05 ## [0.52.3] - 2026-02-05 ## [0.52.2] - 2026-02-05 ## [0.52.1] - 2026-02-05 ## [0.52.0] - 2026-02-05 ## [0.51.6] - 2026-02-04 ## [0.51.5] - 2026-02-04 ## [0.51.4] - 2026-02-03 ## [0.51.3] - 2026-02-03 ## [0.51.2] - 2026-02-03 ## [0.51.1] - 2026-02-02 ## [0.51.0] - 2026-02-01 ## [0.50.9] - 2026-02-01 ## [0.50.8] - 2026-02-01 ## [0.50.7] - 2026-01-31 ## [0.50.6] - 2026-01-30 ## [0.50.5] - 2026-01-30 ## [0.50.3] - 2026-01-29 ## [0.50.2] - 2026-01-29 ## [0.50.1] - 2026-01-26 ## [0.50.0] - 2026-01-26 ## [0.49.3] - 2026-01-22 ## [0.49.2] - 2026-01-19 ## [0.49.1] - 2026-01-18 ## [0.49.0] - 2026-01-17 ## [0.48.0] - 2026-01-16 ## [0.47.0] - 2026-01-16 ## [0.46.0] - 2026-01-15 ## [0.45.7] - 2026-01-13 ## [0.45.6] - 2026-01-13 ## [0.45.5] - 2026-01-13 ## [0.45.4] - 2026-01-13 ## [0.45.3] - 2026-01-13 ## [0.45.2] - 2026-01-13 ## [0.45.1] - 2026-01-13 ## [0.45.0] - 2026-01-13 ## [0.44.0] - 2026-01-12 ## [0.43.0] - 2026-01-11 ## [0.42.5] - 2026-01-11 ### Fixed - Use coding-agent's SessionManager instead of custom MomSessionManager to fix API mismatch crash ([#595](https://github.com/badlogic/pi-mono/issues/595)) ## [0.42.4] - 2026-01-10 ## [0.42.3] - 2026-01-10 ## [0.42.2] - 2026-01-10 ## [0.42.1] - 2026-01-09 ## [0.42.0] - 2026-01-09 ## [0.41.0] - 2026-01-09 ## [0.40.1] - 2026-01-09 ## [0.40.0] - 2026-01-08 ## [0.39.1] - 2026-01-08 ## [0.39.0] - 2026-01-08 ## [0.38.0] - 2026-01-08 ## [0.37.8] - 2026-01-07 ## [0.37.7] - 2026-01-07 ## [0.37.6] - 2026-01-06 ## [0.37.5] - 2026-01-06 ## [0.37.4] - 2026-01-06 ## [0.37.3] - 2026-01-06 ## [0.37.2] - 2026-01-05 ## [0.37.1] - 2026-01-05 ## [0.37.0] - 2026-01-05 ## [0.36.0] - 2026-01-05 ## [0.35.0] - 2026-01-05 ## [0.34.2] - 2026-01-04 ## [0.34.1] - 2026-01-04 ## [0.34.0] - 2026-01-04 ## [0.33.0] - 2026-01-04 ## [0.32.3] - 2026-01-03 ## [0.32.2] - 2026-01-03 ## [0.32.1] - 2026-01-03 ## [0.32.0] - 2026-01-03 ## [0.31.1] - 2026-01-02 ## [0.31.0] - 2026-01-02 ### Breaking Changes - `AgentTool` import moved from `@apholdings/jensen-ai` to `@apholdings/jensen-agent-core` - `AppMessage` type renamed to `AgentMessage` - `Attachment` type replaced with `ImageContent` for image handling - `MomSessionManager.loadSession()` renamed to `buildSessionContex()` - `MomSessionManager.createBranchedSessionFromEntries()` signature changed to `createBranchedSession(leafId)` - `ProviderTransport` removed from Agent config, replaced with direct `getApiKey` callback - `messageTransformer` renamed to `convertToLlm` - `ANTHROPIC_API_KEY`/`ANTHROPIC_OAUTH_TOKEN` no longer checked at startup (deferred to first API call) ### Changed - Session entries now include `id` and `parentId` fields for tree structure support - Auth lookup now uses `AuthStorage` class instead of direct environment variable access - Image attachments use `ImageContent` type with `data` field instead of `Attachment` with `content` - `session.prompt()` now uses `images` option instead of `attachments` ### Added - Support for OAuth login via coding agent's `/login` command (link `~/.pi/agent/auth.json` to `~/.pi/mom/auth.json`) ## [0.20.2] - 2025-12-13 ### Fixed - **Skill paths now use container paths**: Skill file paths in system prompt are translated to container paths (e.g., `/workspace/skills/...`) so mom can read them from inside Docker. ## [0.20.1] - 2025-12-13 ### Added - **Skills auto-discovery**: Mom now automatically discovers skills from `workspace/skills/` and `channel/skills/` directories. Skills are directories containing a `SKILL.md` file with `name` and `description` in YAML frontmatter. Available skills are listed in the system prompt with their descriptions. Mom reads the `SKILL.md` file before using a skill. ## [0.19.2] - 2025-12-12 ### Added - Events system: schedule wake-ups via JSON files in `workspace/events/` - Immediate events: trigger when file is created (for webhooks, external signals) - One-shot events: trigger at specific time (for reminders) - Periodic events: trigger on cron schedule (for recurring tasks) - `SlackBot.enqueueEvent()` for queueing events (max 5 per channel) - `[SILENT]` response marker: deletes status message, posts nothing to Slack (for periodic events with nothing to report) - Events documentation in `docs/events.md` - System prompt section explaining events to mom ## [0.18.8] - 2025-12-12 ### Changed - Timestamp prefix now includes timezone offset (`[YYYY-MM-DD HH:MM:SS+HH:MM]`) ## [0.18.7] - 2025-12-12 ### Added - Timestamp prefix on user messages (`[YYYY-MM-DD HH:MM:SS]`) so mom knows current date/time ### Fixed - Sync deduplication now strips timestamp prefix before comparing ## [0.18.6] - 2025-12-12 ### Fixed - Duplicate message in context when message has attachments (sync from log didn't strip attachment section before comparing) - Use `` delimiter for attachments in messages (easier to parse/strip) ## [0.18.5] - 2025-12-12 ### Added - `--download ` flag to download a channel's full history including thread replies as plain text ### Fixed - Error handling: when agent returns `stopReason: "error"`, main message is updated to "Sorry, something went wrong" and error details are posted to the thread ## [0.18.4] - 2025-12-11 ### Fixed - Attachment downloads now work correctly - SlackBot now receives store for processing file downloads - Files are downloaded in background and stored in `/attachments/` - Attachment paths passed to agent as absolute paths in execution environment - Backfill also downloads attachments from historical messages ## [0.18.3] - 2025-12-11 ### Changed - Complete rewrite of message handling architecture (#115) - Now uses `AgentSession` from coding-agent for session management - Brings auto-compaction, overflow handling, and proper prompt caching - `log.jsonl` is the source of truth for all channel messages - `context.jsonl` stores LLM context (messages sent to Claude, same format as coding-agent) - Sync mechanism ensures context.jsonl stays in sync with log.jsonl at run start - Session header written immediately on new session creation (not lazily) - Tool results preserved in context.jsonl for multi-turn continuity - Backfill improvements - Only backfills channels that already have a `log.jsonl` file - Strips @mentions from backfilled messages (consistent with live messages) - Uses largest timestamp in log for efficient incremental backfill - Fetches DM channels in addition to public/private channels - Message handling improvements - Channel chatter (messages without @mention) logged but doesn't trigger processing - Messages sent while mom is busy are logged and synced on next run - Pre-startup messages (replayed by Slack on reconnect) logged but not auto-processed - Stop command executes immediately (not queued), can interrupt running tasks - Channel @mentions no longer double-logged (was firing both app_mention and message events) - Usage summary now includes context window usage - Shows current context tokens vs model's context window - Example: `Context: 4.2k / 200k (2.1%)` ### Fixed - Slack API errors (msg_too_long) no longer crash the process - Added try/catch error handling to all Slack API calls in the message queue - Main channel messages truncated at 35K with note to ask for elaboration - Thread messages truncated at 20K - replaceMessage also truncated at 35K - Private channel messages not being logged - Added `message.groups` to required bot events in README - Added `groups:history` and `groups:read` to required scopes in README - Stop command now updates "Stopping..." to "Stopped" instead of posting two messages ### Added - Port truncation logic from coding-agent: bash and read tools now use consistent 2000 lines OR 50KB limits with actionable notices ## [0.10.2] - 2025-11-27 ### Breaking Changes - Timestamps now use Slack format (seconds.microseconds) and messages are sorted by `ts` field - **Migration required**: Run `npx tsx scripts/migrate-timestamps.ts ./data` to fix existing logs - Without migration, message context will be incorrectly ordered ### Added - Channel and user ID mappings in system prompt - Fetches all channels bot is member of and all workspace users at startup - Mom can now reference channels by name and mention users properly - Skills documentation in system prompt - Explains custom CLI tools pattern with SKILL.md files - Encourages mom to create reusable tools for recurring tasks - Debug output: writes `last_prompt.txt` to channel directory with full context - Bash working directory info in system prompt (/ for Docker, cwd for host) - Token-efficient log queries that filter out tool calls/results for summaries ### Changed - Turn-based message context instead of raw line count (#68) - Groups consecutive bot messages (tool calls/results) as single turn - "50 turns" now means ~50 conversation exchanges, not 50 log lines - Prevents tool-heavy runs from pushing out conversation context - Messages sorted by Slack timestamp before building context - Fixes out-of-order issues from async attachment downloads - Added monotonic counter for sub-millisecond ordering - Condensed system prompt from ~5k to ~2.7k chars - More concise workspace layout (tree format) - Clearer log query examples (conversation-only vs full details) - Removed redundant guidelines section - User prompt simplified: removed duplicate "Current message" (already in history) - Tool status labels (`_→ label_`) no longer logged to jsonl - Thread messages and thinking no longer double-logged ### Fixed - Duplicate message logging: removed redundant log from app_mention handler - Username obfuscation in thread messages to prevent unwanted pings - Handles @username, bare username, and <@USERID> formats - Escapes special regex characters in usernames ## [0.10.1] - 2025-11-27 ### Changed - Reduced tool verbosity in main Slack messages (#65) - During execution: show tool labels (with → prefix), thinking, and text - After completion: replace main message with only final assistant response - Full audit trail preserved in thread (tool details, thinking, text) - Added promise queue to ensure message updates execute in correct order ## [0.10.0] - 2025-11-27 ### Added - Working memory system with MEMORY.md files - Global workspace memory (`workspace/MEMORY.md`) shared across all channels - Channel-specific memory (`workspace//MEMORY.md`) for per-channel context - Automatic memory loading into system prompt on each request - Mom can update memory files to remember project details, preferences, and context - ISO 8601 date field in log.jsonl for easy date-based grepping - Format: `"date":"2025-11-26T10:44:00.123Z"` - Enables queries like: `grep '"date":"2025-11-26' log.jsonl` - Centralized logging system (`src/log.ts`) - Structured, colored console output (green for user messages, yellow for mom activity, dim for details) - Consistent format: `[HH:MM:SS] [context] message` - Type-safe logging functions for all event types - Usage tracking and cost reporting - Tracks tokens (input, output, cache read, cache write) and costs per run - Displays summary at end of each agent run in console and Slack thread - Example: `💰 Usage: 12,543 in + 847 out (5,234 cache read, 127 cache write) = $0.0234` - Working indicator in Slack messages - Channel messages show "..." while mom is processing - Automatically removed when work completes - Improved stop command behavior - Separate "Stopping..." message that updates to "Stopped" when abort completes - Original working message continues to show tool results (including abort errors) - Clean separation between status and results ### Changed - Enhanced system prompt with clearer directory structure and path examples - Improved memory file path documentation to prevent confusion - Message history format now includes ISO 8601 date for better searchability - System prompt now includes log.jsonl format documentation with grep examples - System prompt now includes current date and time for date-aware operations - Added efficient log query patterns using jq to prevent context overflow - System prompt emphasizes limiting NUMBER of messages (10-50), not truncating message text - Log queries now show full message text and attachments for better context - Fixed jq patterns to handle null/empty attachments with `(.attachments // [])` - Recent messages in system prompt now formatted as TSV (43% token savings vs raw JSONL) - Enhanced security documentation with prompt injection risk warnings and mitigations - **Moved recent messages from system prompt to user message** for better prompt caching - System prompt is now mostly static (only changes when memory files change) - Enables Anthropic's prompt caching to work effectively - Significantly reduces costs on subsequent requests - Switched from Claude Opus 4.5 to Claude Sonnet 4.5 (~40% cost reduction) - Tool result display now extracts actual text instead of showing JSON wrapper - Slack thread messages now show cleaner tool call formatting with duration and label - All console logging centralized and removed from scattered locations - Agent run now returns `{ stopReason }` instead of throwing exceptions - Clean handling of "aborted", "error", "stop", "length", "toolUse" cases - No more error-based control flow ### Fixed - jq query patterns now properly handle messages without attachments (no more errors on empty arrays) ## [0.9.4] - 2025-11-26 ### Added - Initial release of Mom Slack bot - Slack integration with @mentions and DMs - Docker sandbox mode for isolated execution - Bash tool with full shell access - Read, write, edit file tools - Attach tool for sharing files in Slack - Thread-based tool details (clean main messages, verbose details in threads) - Single accumulated message per agent run - Stop command (`@mom stop`) to abort running tasks - Persistent workspace per channel with scratchpad directory - Streaming console output for monitoring