intent(naming): user clarified that Hermes Agent should be written with its formal product name.
decision(naming): updated README and localized Repomix Explorer Skill docs to use Hermes Agent in prose while keeping the hermes CLI command and ~/.hermes path unchanged.
intent(agent-skills): user clarified that Repomix Explorer should use the current npx skills add command and asked to prefer dedicated install flows where supported.
decision(claude-code): documented the Claude Code plugin install path as the recommended Claude Code-specific flow because Repomix ships an official repomix-explorer plugin marketplace entry.
decision(agent-skills): kept npx skills add for Codex, Cursor, OpenClaw, and other Agent Skills-compatible assistants because the Skills CLI targets those agent directories from the official Repomix repository source.
decision(hermes): added Hermes Agent's native hermes skills install command using the raw Repomix Explorer SKILL.md because Hermes documents direct URL skill installs.
The previous commit (71164907) used an incorrect graph token in the
Codecov badge URL, which prevented the badge from rendering. Replace
it with the correct token from codecov.io.
Add JUnit XML reporting to the test-coverage job and upload results
via codecov/test-results-action. This enables Test Analytics features
(test run times, failure rates, flaky test detection, and PR comments
with failed test summaries).
Also update the README Codecov badge URL to include the graph token,
following Codecov's new recommended badge format.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Update all documentation, build configuration, and source comments
to reflect the completed migration from tiktoken to gpt-tokenizer.
- Update tokenCount.encoding description to say "OpenAI-compatible
tokenization" with gpt-tokenizer link (README + 15 language docs)
- Remove tiktoken from bundling external deps list since
gpt-tokenizer is pure JS (README + 15 language docs)
- Remove tiktoken COPY from Dockerfile and external from bundle.mjs
- Simplify source code comments removing migration artifacts
- Regenerate scripts/memory/package-lock.json
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Tower (git-tower) cancelled their sponsorship, so removing their logo,
links, and image assets from README and website.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Community projects are inherently community-maintained, so the
"unofficial" qualifier is redundant. Remove it from README and all
15 language versions of community-projects.md.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Add CLI option reference and a security note explaining that remote
repository config files are not loaded by default, with instructions
for opting in via --remote-trust-config or REPOMIX_REMOTE_TRUST_CONFIG.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Document the ready-to-use Repomix Explorer skill that enables AI coding
assistants to analyze and explore codebases. Include quick install command
using add-skill and describe what the skill enables.
Change "for code compression feature" to "required for code compression feature"
for consistency with other documentation pages as suggested in PR review.
Document --split-output feature with:
- Use case (AI Studio's 1MB file size limit)
- Usage example and file naming convention
- Note about directory-based grouping
- Add 'gitignore' language to .gitignore example in TIP block
- Add 'text' language to directory structure examples
- Fix nested fenced code blocks in files.md section (use 4 backticks)
- Update kebab-case example to use 'CustomName' instead of 'custom-name'
Applied to: en, de, es, fr, hi, id
- Add explanation that this feature is particularly powerful for
referencing implementations from remote repositories
- Remove redundant "What is Agent Skills Generation?" section
- Merge explanation into the introduction paragraph
Update all skill-generate documentation to use "Agent Skills" or "Skills"
instead of "skill" for consistency:
- README.md: Update section titles and content
- English website doc: Update to "Agent Skills Generation"
- Japanese website doc: Update to "Agent Skills生成"
- Sidebar labels: Update both English and Japanese labels
Add documentation for the --skill-generate feature introduced in PR #998:
- Add --skill-generate option to Command Line Options section
- Add dedicated Skill Generation section with usage examples
- Document generated structure (SKILL.md and references/)
- Include auto-generated skill names and integration examples
Add a dedicated Privacy Policy page to the website, addressing user
feedback about privacy information being hard to find on repomix.com.
Changes:
- Create Privacy Policy page with content from README
- Add sidebar links for all supported languages
- Non-English pages include English content via @include
- Simplify README Privacy Policy section to link only
Related Discord feedback requested clearer privacy documentation
on the website for users who need to verify data handling before
using Repomix in their work environment.
Updated priority order documentation in README.md and all website configuration guides (12 languages) to accurately reflect how globby handles ignore files:
- Merged items 2-4 into a single "Ignore files" category
- Clarified that directory hierarchy determines priority (deeper directories win)
- Explained that files in the same directory are merged in no particular order
- This reflects globby's actual behavior where fast-glob returns files in filesystem order
The previous documentation incorrectly implied a fixed priority order between .repomixignore, .ignore, and .gitignore files, when in fact they are all merged together at the same level with priority determined by directory depth only.
This PR adds support for .ignore files, which are used by tools like ripgrep and the silver searcher. This allows users to maintain a single .ignore file that works across multiple tools instead of maintaining separate ignore files.
Changes:
- Add ignore.useDotIgnore config option (default: true)
- Add --no-dot-ignore CLI flag to disable .ignore file usage
- Update ignore file priority: .repomixignore > .ignore > .gitignore > default patterns
- Add comprehensive tests for .ignore file handling
- Update documentation to reflect new .ignore file support
The .ignore file is enabled by default but can be disabled via configuration or CLI flag, maintaining backward compatibility.
Resolves#937
Renamed the repository explorer plugin from `repository-explorer` to
`repomix-explorer` for better alignment with the Repomix ecosystem naming
convention. Also simplified the agent name from `repository-explorer` to
`explorer` for brevity.
Changes:
- Plugin name: `repository-explorer` → `repomix-explorer`
- Agent name: `repository-explorer` → `explorer`
- Fully qualified name: `repository-explorer:repository-explorer` → `repomix-explorer:explorer`
This change affects:
- Plugin directory (.claude/plugins/repository-explorer → repomix-explorer)
- Agent file (agents/repository-explorer.md → agents/explorer.md)
- plugin.json metadata
- All command files (explore-local.md, explore-remote.md)
- repomix-commands plugin references
- marketplace.json
- README.md
- Website documentation (12 language versions)
The new naming provides:
1. Better consistency with other Repomix plugins (repomix-mcp, repomix-commands)
2. Clearer brand association with Repomix
3. More concise agent name that's easier to reference
BREAKING CHANGE: Users need to update their plugin installation:
- Old: `/plugin install repository-explorer@repomix`
- New: `/plugin install repomix-explorer@repomix`
Commands are also updated:
- Old: `/repository-explorer:explore-local`, `/repository-explorer:explore-remote`
- New: `/repomix-explorer:explore-local`, `/repomix-explorer:explore-remote`
Updated all documentation to use the correct fully qualified agent name
format (plugin-name:agent-name) for the repository-explorer agent.
Changed from:
- repository-explorer
To:
- repository-explorer:repository-explorer
This change ensures consistency with Claude Code's agent naming convention
where agents are referenced using the format plugin-name:agent-name when
the plugin name and agent name are the same.
Files updated:
- README.md
- Plugin documentation (6 files in .claude/plugins/)
- Website documentation (12 language versions)
Add comprehensive documentation for the new repository-explorer plugin
across README and website documentation (English and Japanese versions).
Changes:
- README.md: Add repository-explorer as the 3rd plugin with features,
commands, and usage examples
- Update installation section with repository-explorer install command
- Update plugin relationship note to mention all three plugins
Website documentation (en/ja):
- Add repository-explorer plugin section with features and commands
- Add AI-powered analysis usage examples for local and remote exploration
- Update installation instructions and plugin relationship tips
- Add repository-explorer to plugin source code links
Other changes:
- Bump marketplace.json version to 1.0.2
- Bump repomix-commands plugin version to 1.0.2
The repository-explorer plugin provides AI-powered codebase analysis
that intelligently explores repositories using natural language commands
while managing context efficiently through incremental grep and read
operations.
Add CLI option documentation for the --include-full-directory-structure flag
in the README. This flag allows users to display the complete directory tree
in output, including files not matched by --include patterns, providing full
project context while maintaining focused file selection.
- Update example repository name to be consistent
- Clarify plugin installation order and relationship
- Add 'text' language tag to code blocks
- Explain recommended plugin installation process
Addresses review comments from Gemini Code Assist and Copilot
Move the Claude Code Plugins section after the Available MCP Tools section for better documentation flow. This creates a more logical structure:
1. MCP Server Integration (overview)
2. Available MCP Tools (technical details)
3. Claude Code Plugins (user-friendly integration)
4. Configuration
This organization helps users understand the MCP foundation before learning about the convenient Claude Code plugin wrappers.
Update command examples to use the actual format with plugin prefix:
- /pack-local -> /repomix-commands:pack-local
- /pack-remote -> /repomix-commands:pack-remote
This matches the actual command format shown in Claude Code.
Add comprehensive documentation for Claude Code plugins:
- Overview of repomix-mcp and repomix-commands plugins
- Installation instructions for Claude Code
- Example usage with natural language commands
- Benefits and features of using plugins
- Cross-reference from MCP Server Integration section
This provides users with clear guidance on how to use Repomix with Claude Code through the official plugin system.
Added $schema property to the example configuration to demonstrate best
practice and encourage users to enable schema validation for better IDE
support.
Added documentation for JSON schema validation feature that was present in
the website docs but missing from README. Users can now enable auto-completion
and validation in their editors by adding the $schema property to their
configuration file.
Removed 'Real-time error detection while editing' as it's redundant with
'Full TypeScript type checking in your IDE' which already implies real-time
error detection.
Reduced the JavaScript Configuration section to a single line noting it works
the same as TypeScript. Removed redundant examples since the behavior and
usage are identical to TypeScript configuration.
Kazuki Yamada
Dango233
Claude