1 Version 1.32.1
════════════════

  • Preserve file buffer positions during tool updates. File contents
    are now updated with marker-preserving replacement, and only stale
    unmodified buffers are refreshed after shell commands and edit
    hooks. The system warns instead of discarding unsaved changes, and
    regression coverage for mutating, read-only, validation, and hook
    paths has been added.

  • Keep visiting buffers attached after file moves. The `move_file'
    tool now retargets an existing source buffer to the destination
    while preserving point and unsaved changes. Destinations already
    visited by another buffer are rejected. Regression tests cover both
    paths.


2 Version 1.32.0
════════════════

  • Preserve Markdown tables during streaming conversion. Paragraph
    filling no longer wraps table rows, and streaming fill skips while
    point is inside a table, avoiding phantom rows and broken cells.
  • Align converted Markdown tables in Org output. Separator rows are
    normalized and tables are aligned for both builtin and Pandoc
    converters, with better handling of wider cells during streaming
    updates.


3 Version 1.31.0
════════════════

  • Show irreversible action confirmation details. Named tool arguments
    are now included in security confirmation prompts, and a full-detail
    view flow has been added for DLP and irreversible actions. The flow
    preserves context even when audit logging fails, and regression
    tests cover prompt contents as well as the view interaction.


4 Version 1.30.0
════════════════

  • Split the subagent task tool into two separate contracts. The `task'
    tool handles free-form delegation, while `task_from_template'
    handles template-based delegation. Both now have proper required
    schemas and validation checks instead of accepting anything
    silently. The direct Lisp template interface still works as
    before. Updated the docs and added regression tests.
  • Added a plan-and-act action to the blueprint-mode transient
    menu. Starting it kicks off the plan-and-act agent loop from the
    current blueprint buffer and then closes the buffer. Included
    regression tests to verify the buffer contents get forwarded to
    `ellama-plan-and-act' and that the source buffer is killed once the
    loop starts.


5 Version 1.29.0
════════════════

  • Added `ellama-setup-agentic-coding' to configure long-running
    coding-agent defaults. This profile enables DLP enforcement,
    fail-closed DLP behavior, session compaction, visible sub-agent
    sessions, and longer agent loop limits. In non-SRT mode, ordinary
    DLP findings ask first while normal confirmations stay active; in
    SRT mode, ordinary input findings are allowed and ordinary output
    findings are redacted, with irreversible findings still
    blocked. Tests cover non-SRT, SRT, and secure modes.
  • Removed obsolete problem solver commands
    `ellama-solve-reasoning-problem' and
    `ellama-solve-domain-specific-problem' from the codebase and the
    main transient menu, along with the `ellama-chain' function they
    relied upon.
  • Updated the README onboarding section to refresh Installation
    references around GNU ELPA, Ollama, and added an explicit
    `qwen3.6:35b' agentic setup example.
  • Documented the new `ellama-setup-agentic-coding' profile and split
    Commands and Configuration sections into smaller groups for easier
    navigation. Regenerated `ellama.info'.


6 Version 1.28.0
════════════════

  • Audio input support: audio files can now be attached alongside
    images in prompts, with provider capability validation. Includes
    transient entries, key bindings, a recording lighter in the mode
    line, and microphone recording commands. `ellama-read-file' can
    queue audio files for the next model turn, exposing audio mode in
    customization and tool metadata.
  • Audio recording backends: on startup, auto-selects the available
    recorder by platform — FFmpeg or SoX on macOS, arecord/FFmpeg/SoX on
    GNU/Linux. Fallback failures include the recorder's output and a
    macOS microphone-permission hint.
  • Recording normalization: default FFmpeg recordings are processed
    through a configurable dynamic audio normalization filter so quiet
    speech reaches audio
  …  …
