* chore(nix): improve flake with dynamic version and source filtering
- Read version dynamically from package.json instead of hardcoding
- Add lib.fileset source filtering to exclude node_modules and build artifacts
- Update update-flake.sh to support dynamic version pattern
- Add hash change detection to skip unnecessary rebuilds
- Improve error handling with automatic rollback on failure
- Update specs to reflect dynamic version behavior
* chore(ci): bump Nix actions to latest versions
- nix-installer-action: v13 → v21
- magic-nix-cache-action: v8 → v13
- Update validation message for unchanged flake.nix
* chore: add changeset for Nix improvements
* fix(nix): make update-flake.sh portable to macOS
- Fix grep pattern on line 37 to include opening parenthesis
- Replace GNU grep -oP with portable sed alternatives (lines 53, 68, 70)
- Ensures script works on both Linux and macOS (BSD sed/grep)
* fix(nix): properly check build verification exit status
Fix logic bug where build failures were incorrectly reported as success.
The script now:
- Captures build exit code and output separately
- Fails fast if build returns non-zero exit code
- Only checks for 'dirty tree' warning if build succeeded
This addresses CodeRabbit review feedback on line 101-107.
---------
Co-authored-by: Tabish Bidiwale <30385142+TabishB@users.noreply.github.com>
* feat(skills): add opsx:onboard guided workflow skill
Add a new onboard skill that walks users through their first complete
OpenSpec workflow cycle. The skill provides interactive guidance through
task selection, change creation, artifact building, implementation, and
archiving.
Also includes:
- New README with updated branding and workflow examples
- Documentation structure placeholders
- Change artifacts for the onboard skill feature
* test(skills): update skill-generation tests for onboard skill
Update test expectations from 9 to 10 skills after adding opsx:onboard.
* docs: update README links and add doc cleanup checklist
- Replace placeholder links in README_NEW.md with actual doc paths
- Add documentation cleanup checklist to README_RENEWAL_PROMPTS.md
* docs: overhaul documentation with new workflows, getting-started, and customization guides
- Rewrite workflows.md with action-based philosophy and workflow patterns
- Rewrite getting-started.md with clearer onboarding flow
- Rewrite customization.md with schema customization guidance
- Add cross-references between docs (Commands, Customization links)
- Remove obsolete docs: artifact_poc, experimental-release-plan, project-config-demo, schema-customization, schema-workflow-gaps
- Update README_RENEWAL_PROMPTS.md checklist
* docs: continue documentation overhaul with expanded guides and restructuring
- Expand cli.md, commands.md, and concepts.md with comprehensive content
- Add installation.md, multi-language.md, and supported-tools.md
- Rename experimental-workflow.md to opsx.md
- Remove i18n.md (replaced by multi-language.md)
- Update README links and cleanup prompts
* chore(assets): consolidate logo images
* docs: enhance README with badges, usage notes, and contributing guidelines
- Update Discord badge to show member count
- Add collapsible section with stars/downloads/contributors badges
- Add OpenSpec Dashboard preview section
- Add usage notes for model selection and context hygiene
- Expand contributing section with guidelines for small/large changes
- Clarify AI-generated code policy
* docs: remove misleading mid-flight update claims
The documentation claimed users could edit artifacts mid-implementation
and seamlessly continue, but no such mechanism exists. This removes:
- "Mid-Flight Correction" section from workflows.md
- Feedback arrows and "update as you learn" from all diagrams
- Mid-flight claims from commands.md, opsx.md, concepts.md
- Example blocks showing edit-then-continue workflow
Also adds a proposal for future artifact regeneration support that
would actually make this workflow possible.
* docs: fix PR review comments (markdown linting and accuracy)
- Add language tags to fenced code blocks (MD040)
- Remove blank line between blockquotes (MD028)
- Capitalize "Markdown" as proper noun
- Update deprecated command reference (experimental -> update)
- Update skill count from 9 to 10, add openspec-onboard
- Fix typo: fix-midlight -> fix-midflight
* chore: remove polish-release-notes CI workflow
Replaced with local /polish-release skill. The claude-code-action
doesn't work well with repository_dispatch triggers (no PR context).
* docs: clarify /opsx:sync is optional (archive prompts if needed)
Remove sync from main workflow flows and diagrams since archive
already prompts to sync when needed. Most users will never need
to call sync directly.
- Remove sync from completion flow diagrams
- Remove "Sync Specs Regularly" best practice section
- Update command descriptions to note it's optional
- Update "When to sync" to "When to use manually"
* docs: redesign README with simplified content and new OPSX callout
- Simplify badges and logo presentation
- Add collapsible "most loved" section
- Replace detailed explanation with concise philosophy
- Add prominent /opsx:onboard callout for new workflow
- Remove README_NEW.md (content merged into README.md)
- Remove renewal prompts documentation
* docs: add README_OLD.md as reference backup
* docs: fix command directory paths for multiple tools
Correct commands locations for Antigravity, Codex, Crush, OpenCode,
and Qoder in the supported tools table.
The GitHub App token doesn't have actions:write permission, which is
required for workflow_dispatch. Switch to repository_dispatch which
works with existing contents:write permission.
Changes:
- release-prepare.yml: Use gh api to trigger repository_dispatch
- polish-release-notes.yml: Add repository_dispatch trigger type
- Delete test workflow (validation complete)
Tested via PR #542 - both workflow_dispatch and repository_dispatch
triggers work correctly with claude-code-action.
* fix(ci): use workflow_dispatch for polish release notes
The claude-code-action doesn't support the `release` event type.
Switch to workflow_dispatch which is supported, and have
release-prepare trigger it after publishing.
* fix: get tag from package.json instead of gh release list
* perf: add path filtering to Nix validation CI job
Skip Nix flake validation when no Nix-related files change. This speeds
up CI for PRs that only touch source code, docs, or tests.
Files that trigger Nix validation:
- flake.nix, flake.lock
- package.json, pnpm-lock.yaml
- scripts/update-flake.sh
- .github/workflows/ci.yml
Uses dorny/paths-filter@v3 for change detection. Required-checks jobs
updated to handle skipped status correctly.
* fix: auto-trigger polish release notes on release publish
The workflow was only set up for manual dispatch, requiring someone to
remember to run it after each release. This change adds an automatic
trigger on `release: [published]` events while keeping the manual
trigger as a fallback.
The Claude Code Action supports any GitHub event when using the `prompt`
parameter for custom automations.
Skip Nix flake validation when no Nix-related files change. This speeds
up CI for PRs that only touch source code, docs, or tests.
Files that trigger Nix validation:
- flake.nix, flake.lock
- package.json, pnpm-lock.yaml
- scripts/update-flake.sh
- .github/workflows/ci.yml
Uses dorny/paths-filter@v3 for change detection. Required-checks jobs
updated to handle skipped status correctly.
The workflow was failing because Claude Code requested permission to write files
(release-title.txt and polished-notes.md) but there was no interactive user to approve.
Added claude_args: "--allowedTools Write,Read" to pre-approve file operations
in automation mode.
The workflow was failing due to two issues:
1. Using unsupported 'release' event trigger - Claude Code action doesn't support this event type
2. Using deprecated 'direct_prompt' parameter instead of 'prompt'
Changes:
- Switch from 'release' trigger to 'workflow_dispatch' for manual triggering
- Replace 'direct_prompt' with 'prompt' parameter (v1.0 breaking change)
- Update all tag references from github.event.release.tag_name to inputs.tag_name
The workflow now needs to be manually triggered from GitHub Actions UI after a release is published.
GitHub's auto-merge feature uses an internal token that cannot trigger
workflows. Even with manual approval, the merge performed by auto-merge
doesn't trigger the release workflow.
Removing auto-merge means:
1. Version PR is created with CI running
2. User manually merges the PR (clicking "Merge")
3. The merge triggers the release workflow
4. Package is published
This aligns with industry standard practice for changesets.
Move the GitHub App token generation before checkout and pass it
to actions/checkout. This configures git to use the App's identity
for all operations, allowing pushes to trigger CI workflows.
Removes commitMode: github-api which doesn't support executable files.
* ci: use GitHub App token to trigger CI on version PR
Replace GITHUB_TOKEN with a GitHub App token so that the version PR
can trigger CI workflows. GITHUB_TOKEN cannot trigger workflows by
design (to prevent infinite loops).
Requires APP_ID variable and APP_PRIVATE_KEY secret to be configured.
* ci: upgrade create-github-app-token to v2
* ci: add commitMode github-api to trigger CI on version PR
* ci: use GitHub App token to trigger CI on version PR
Replace GITHUB_TOKEN with a GitHub App token so that the version PR
can trigger CI workflows. GITHUB_TOKEN cannot trigger workflows by
design (to prevent infinite loops).
Requires APP_ID variable and APP_PRIVATE_KEY secret to be configured.
* ci: upgrade create-github-app-token to v2
* Add changeset for Continue support, shell completions, and explore command
* ci: auto-merge version PR to streamline releases
Enable auto-merge on the changesets version PR so it merges
automatically once CI passes. This reduces the release process
from 2 manual PR merges to effectively 1.
Requires enabling "Allow auto-merge" in repository settings
and branch protection rules on main.
* fix(cli): use dynamic import for @inquirer/prompts in config command
The config command (added in #382) reintroduced the pre-commit hook hang
issue that was fixed in #380. The static import of @inquirer/prompts at
module load time causes stdin event listeners to be registered even when
running non-interactive commands, preventing clean process exit when
stdin is piped (as pre-commit does).
Convert the static import to a dynamic import that only loads inquirer
when the `config reset` command is actually used interactively.
Fixes#367
* chore: add ESLint with no-restricted-imports rule for @inquirer
Add ESLint configuration that prevents static imports of @inquirer/*
modules. This prevents future regressions of the pre-commit hook hang
issue fixed in this PR.
The rule shows a helpful error message pointing to issue #367 for context.
init.ts is exempted since it's already dynamically imported from the CLI.
* ci: add ESLint step to lint job
Run `pnpm lint` in CI to enforce the no-restricted-imports rule
that prevents static @inquirer imports.
* fix: correct CLI version mismatch and add release guard\n\n- Add patch changeset for 0.8.1\n- Add pack-version check to validate tarball version\n- Update release script to include versioning and guard
* chore(release): simplify release script and harden pack-version-check\n\n- Run pack-version check before publish only\n- Remove redundant changeset version + explicit build in release script\n- Always cleanup temp dir and tgz\n- Quieter, faster npm install during guard
* chore(release): clarify CI vs local release scripts; refine pack guard\n\n- Add scripts: release:ci (no version), release:local (runs changeset version)\n- Workflow uses release:ci to ensure version PR bump precedes publish\n- Pack guard: document npm vs pnpm choice; improve JSON fallback handling
* feat: implement Phase 1 E2E testing with cross-platform CI matrix
- Add shared runCLI helper in test/helpers/run-cli.ts for spawn testing
- Create test/cli-e2e/basic.test.ts covering help, version, validate flows
- Migrate existing CLI exec tests to use runCLI helper
- Extend CI matrix to bash (Linux/macOS) and pwsh (Windows)
- Update Phase 1 tasks and proposal with implementation status
* fix: correct YAML syntax in CI workflow diagnostics command
* fix: use multiline YAML for diagnostics command
* fix ci
* fix: ci
* fix: update core validation and json converter
* chore(ci): split pr and main workflows
* refactor: simplify CI workflow with unified matrix strategy
- Consolidate test_pr and test_matrix into single test job
- Add proper shell configuration with defaults
- Add timeout protection (15 minutes)
- Simplify required-checks to single job
- Maintain cross-platform testing (bash on Linux/macOS, pwsh on Windows)
* fix: restore lean PR workflow with async main branch matrix
- PRs run only essential tests on ubuntu-latest (fast feedback)
- Main branch runs full cross-platform matrix asynchronously
- Separate required-checks for each workflow type
- Different timeouts: 10min for PR, 15min for matrix
