v0.2.11

# Release Notes for v0.2.11

This release adds **real write-time contradiction detection and resolution** (replacing the "MVP direct store" stub with an actual `MemoryValidationService` and batch-prefetch performance path), plus a very large batch of validation and correctness fixes across config management, CLI connect/skill install, MCP session isolation, and the LangGraph/Hermes/CrewAI integrations.

## New Features

- **Write-time memory contradiction resolution** (`memanto/app/services/memory_validation_service.py`,
  `memanto/app/services/memory_write_service.py`)
  - Write-time validation was previously stubbed out ("MVP direct store"), so
    logically contradictory memories were stored blindly. New
    `MemoryValidationService` detects contradictions deterministically (no LLM
    on the write path): a stored *active* memory with the same type and title
    (case-insensitive) but different content is a contradiction.
  - Resolution follows the `keep_new` convention already used by the offline
    conflict report: the old record is preserved as history
    (`status=superseded`, `superseded_by`/`superseded_at` — the same fields
    `search_as_of` already reads) via the same delete-and-recreate pattern
    `update_memory` uses; the new record is stored active.
  - Contradictions within a single batch resolve last-wins, with earlier
    conflicting entries stored as superseded history.
  - Only memory types in `settings.REQUIRE_VALIDATION_FOR` are checked, keeping
    the write path cheap for types that don't need it.
  - **Batch prefetching**: contradiction candidates for an entire batch are
    fetched together up front instead of one Moorcheh lookup per memory,
    substantially reducing write latency for batch operations.
  - The supersede step now uploads the replacement document directly (instead
    of a separate update call) and no longer deletes the original record if
    the upload fails — a failed supersede leaves the original intact rather
    than losing data.
  - Manual conflict resolution now stores the replacement content **before**
    deleting the superseded record, and a failed deletion during resolution
    correctly leaves the conflict marked unresolved instead of silently
    dropping it.

## Improvements

- **Config validation hardening** (`memanto/cli/config/manager.py`,
  `memanto/app/ui/routes/ui_router.py`, `memanto/cli/schedule_time.py`)
  - Server URLs are normalized and ports validated before being written to
    config; session/answer config edits are validated against their schema;
    config setters now guard against malformed config sections instead of
    crashing; schedule times are validated (format + range) before a
    scheduled job is enabled, both via the CLI and the UI (invalid times now
    return 400 instead of silently accepting garbage).

- **Moorcheh client cache invalidation** (`memanto/app/clients/moorcheh.py`)
  - `MoorchehClientSingleton` now tracks the config it was built from
    (backend, URL, timeout / API key) and rebuilds the cached client when
    that config changes, instead of serving a stale client after a backend
    or API-key switch.

- **Avoid storage writes during service initialization** (`memanto/app/services/session_service.py`,
  `memanto/app/services/agent_service.py`)
  - `SessionService`/`AgentService` no longer eagerly create directories or
    generate a secret key at construction time; both are now lazy
    (created/generated on first actual use), avoiding unnecessary filesystem
    writes (and directory creation) just from instantiating the service.

- **`memanto connect` / skill install correctness** (`memanto/cli/connect/engine.py`,
  `memanto/cli/connect/agent_registry.py`)
  - Fixed false-positive shared-skill detection that could report a skill as
    installed for one agent when it was actually another agent's install.
  - `connect --disconnect` now preserves unmanaged rule files (including
    non-UTF-8 ones) instead of deleting content Memanto didn't add.
  - Hook-cleanup matching (used when disconnecting) tightened to avoid
    removing unrelated hook entries.

- **Agent list namespace counts hardened** (`memanto/cli/config/manager.py`)
  - `list agents` now guards against malformed/missing `item_count` or
    `namespace_name` fields in the Moorcheh namespaces response instead of
    raising.

- **MCP integration: per-agent session isolation** (`integrations/mcp/memanto_mcp/lifecycle.py`,
  `integrations/mcp/memanto_mcp/tools.py`)
  - Concurrent MCP tool calls for different agent IDs no longer share one
    `SdkClient` session — each agent gets its own isolated, cached client, with
    batch payload validation happening before any readiness side effects and
    newly created scoped clients only cached once activation actually succeeds.
  - MCP `batch_remember` now guards against and normalizes malformed item
    results instead of propagating a bad shape.

- **LangGraph integration fixes** (`integrations/langgraph/langgraph_memanto/{store,tools}.py`)
  - Non-string content values are stringified before being written to the
    store (previously could raise). `min_confidence` search filtering fixed.
    Setup now retries after an activation failure instead of getting stuck.
    Input limits enforced on `remember`.

- **Hermes / CrewAI integration fixes** (`integrations/hermes-agents/`, `integrations/crewai/`)
  - Fixed agent-id truncation collisions in Hermes (two different agent IDs
    could truncate to the same internal id). Hermes memory-mirror writes on
    exit are now non-blocking. CrewAI `remember` now enforces the same input
    length limits as the core API.

- **Memory validation / parsing hardening** (`memanto/app/services/memory_parsing_service.py`,
  `memanto/app/models/__init__.py`)
  - Fixed the ambiguity guard being systemically bypassed by common auxiliary
    verbs (`is`/`are`/`was`/`were`) matching `STRONG_FACT_PATTERNS`.
  - Memory type is now schema-validated on write instead of accepted as any
    string; `SourceType` validation is strict and rate-limiting now fails
    closed on error instead of open.
  - Tags and source labels are now length-bounded before storage (applies to
    both `remember` and memory edits) using shared Pydantic
    `Annotated[StringConstraints]` validators.
  - Fail-fast on malformed storage responses in the core write path,
    `memanto migrate`, and MCP `batch_remember`.

- **Timestamp formatting fix** (`memanto/app/services/memory_read_service.py`)
  - Epoch-integer timestamps are now coerced to ISO strings during memory-read
    formatting instead of being returned as raw ints, which broke downstream
    date parsing.

- **Unicode agent-ID isolation** — agent ID validation is now centralized and
  Unicode-normalized, fixing a case where visually-similar Unicode agent IDs
  could collide or bypass isolation.

## Tests

- New `tests/test_contradiction.py`, `tests/test_conflict_manual_resolution.py`,
  `tests/test_conflict_resolution_delete_failures.py`, `tests/test_connect_engine.py`,
  `tests/test_connect_detection.py`, `tests/test_config_manager_setters.py`,
  `tests/test_schedule_time_validation.py`, `integrations/mcp/tests/test_lifecycle.py`.
- Large expansion of `tests/test_api.py`, `tests/test_backend.py`,
  `tests/test_unit.py`, and per-integration test suites covering every fix
  above.

## Full Changelog

Full Changelog: https://github.com/moorcheh-ai/memanto/compare/v0.2.10...v0.2.11