Skip to content

Releases: zazencodes/aicmd

v0.3.0

Choose a tag to compare

@github-actions github-actions released this 23 Aug 20:15
4711181

What's new

Nothing changes for you. This release is an internal restructure with no
user-facing changes: the same flags, the same config keys, the same
providers, the same output.

Upgrading is optional. If you do upgrade, aicmd --version will report
v0.3.0 and everything else behaves exactly as v0.2.0 did — verified by
diffing the two binaries' output across the whole CLI surface, including
--help, every config subcommand, error paths, and end-to-end command
generation.

Why the release

Go source moved out of the repository root into internal/ packages
(cli, config, provider, terminal, debug), leaving one .go file
at the root. This is groundwork for future provider work; it ships as its
own version so that any behaviour change in a later release can be bisected
against a known-identical baseline.

Install and upgrade steps are unchanged:

curl -fsSL https://github.com/zazencodes/aicmd/raw/refs/heads/main/install.sh | sh

v0.2.0

Choose a tag to compare

@github-actions github-actions released this 22 Aug 19:05
  • Publish concise, curated release notes from annotated version tags instead
    of generated commit lists.
  • Use GitHub's current branch download URL for the README install command to
    avoid stale raw-content caches.

The command-generation behavior and supported release platforms are unchanged
from v0.1.1.

v0.1.1

Choose a tag to compare

@github-actions github-actions released this 22 Aug 18:41

Fixed

  • Fix anonymous downloads in install.sh by using stable GitHub release asset URLs instead of parsing API JSON.

Changed

  • Update the GitHub Actions runtimes used by CI and releases.

v0.1.0

Choose a tag to compare

@github-actions github-actions released this 22 Aug 18:38

Initial public release of aicmd.

Highlights

  • Generate shell commands through Codex, Claude Code, Antigravity, OpenAI, Gemini, Anthropic, Ollama, or llama.cpp.
  • Select providers, models, and reasoning effort with one provider/model/effort spec.
  • Insert generated commands at the terminal prompt for review, with a safe print fallback where terminal injection is unavailable.
  • Manage model and clipboard settings with aicmd config.
  • Stream provider reasoning and raw events in verbose mode.
  • Install checksum-verified binaries for macOS and Linux on amd64 and arm64.

Safety

Prompts are sent to provider CLIs through stdin. Generated output is cleaned and presented for review; aicmd never executes it automatically.