Releases: imainzed5/openprompting
Release list
openPrompting v1.1.2
openPrompting v1.1.1
openPrompting 1.1.1
openPrompting 1.1.1 is a documentation-only patch release in the additive V1.1 line.
Fixed
- Updated the README to identify the current public release as
openprompting@1.1.1. - Added the v1.1.1 release-notes link and clarified that the v1.1.0 catalog expansion remains the foundation of this patch release.
Compatibility
The executable remains openprompting, and the command surface remains help, setup, guide, new, doctor, and compare. Configuration continues to use version: 1; existing V1 and V1.1 configuration values and IDs remain valid. Normal CLI operation remains local and deterministic.
openPrompting v1.1.0
openPrompting 1.1.0
openPrompting 1.1.0 expands the source-backed catalog while keeping the V1 local CLI and configuration contracts stable.
Highlights
- Added sourced model guidance for OpenAI GPT-5.3-Codex and Anthropic Claude Opus 5.
- Added harness guidance for Cursor, GitHub Copilot CLI, OpenCode, and Aider, with model behavior kept separate from harness behavior.
- Added planning, research, testing, and migration task recipes with matching deterministic templates.
- Added task/template integrity validation, dynamic catalog checks, and packed-package journeys for every active task.
- Added optional
display_namemetadata so setup, guide, and compare output can be easier to scan without changing stable IDs. - Deepened current GPT-6 Astra, Claude Sonnet 5, and Gemini 3.8 Flash guidance from first-party documentation.
Compatibility
The executable remains openprompting, and the command surface remains help, setup, guide, new, doctor, and compare. Configuration continues to use version: 1; existing V1 configuration values and IDs resolve unchanged. The evidence vocabulary remains official, tested, community, and legacy.
Validation
The release gate covers lint, type checking, the full test suite, build, knowledge validation, README verification, package dry-run, and installation of the packed artifact into an isolated consumer. The packed consumer exercises the original V1 journey, new model and harness selectors, and both default and explicit-profile rendering for every active task.
Operation and privacy
Normal CLI operation remains local and deterministic. This release adds no runtime model calls, telemetry, hosted backend, workflow execution, or automatic network access.
Deferred
Remote knowledge updates, native harness adapters, handoffs and workflows, prompt optimization, hosted documentation, benchmarking, and additional harnesses remain outside this release.
openPrompting v1.0.0
openPrompting 1.0.0
The first stable release delivers a local, deterministic CLI for source-backed model, harness, and task guidance.
Highlights
- Configure one or more model/harness setups, named profiles, defaults, and task routes.
- Resolve guidance and generate structured prompt skeletons for five recurring task types.
- Diagnose local configuration, routing, instruction files, declared scripts, and guidance freshness without executing project code.
- Compare profiles, setups, models, and harnesses without inventing a universal ranking.
- Ship verified initial coverage for OpenAI/Codex, Anthropic/Claude Code, and Google/Gemini CLI.
- Validate knowledge, schemas, mixed-model routing, package contents, and packed-artifact execution in release gates.
Stable contracts
The executable, six command names, config schema v1, evidence vocabulary, and key data directories are frozen. See V1 contracts and migration notes.
Privacy and operation
Normal operation requires no account, hosted service, API key, model call, telemetry, or network connection. Setup writes only its project-local config; doctor is read-only.
Deferred beyond V1
Hosted documentation, native harness adapters, handoffs/workflows, prompt optimization, remote knowledge updates, broader harness coverage, and benchmarking remain in the post-V1 backlog.
openPrompting v1.0.0-rc.1
openPrompting 1.0.0-rc.1
The V1 release candidate froze the executable name, six commands, config schema v1, evidence vocabulary, and primary package directories for installation testing.
Candidate scope
- Completed the three intended initial ecosystems: OpenAI/Codex, Anthropic/Claude Code, and Google/Gemini CLI.
- Added contributor templates, evidence rules, source verification, stale-guidance documentation, and release notes.
- Added cross-platform CI and an isolated packed-artifact journey covering beginner and mixed-model use.
- Audited the package allowlist and migration guidance.
No product features were added after the candidate boundary. The stable 1.0.0 promotion contains release-hardening and documentation corrections only.