Simplify CLI: Accept URLs directly for indexing by richhankins · Pull Request #12 · augmentcode/context-connectors

richhankins · 2026-01-30T08:48:30Z

Summary

Simplifies the ctxc index command to accept URLs directly, eliminating the need for verbose source-specific flags.

Before

ctxc index github --owner augmentcode --repo context-connectors -i context-connectors
ctxc index gitlab --project group/project -i my-project
ctxc index website --url https://docs.example.com -i my-docs

After

ctxc index https://github.com/augmentcode/context-connectors
ctxc index https://gitlab.com/group/project
ctxc index https://docs.example.com

Changes

1. URL Parser Module (`src/core/url-parser.ts`)

New parseSourceUrl() function that auto-detects source type from URLs
Supports GitHub, GitLab, Bitbucket, and website URLs
Extracts owner/repo/project from URL paths
Detects branch/tag from URLs like /tree/main
19 unit tests

2. CLI URL Mode (`src/bin/cmd-index.ts`)

New url subcommand: ctxc index url <url>
Auto-detects source type and creates appropriate indexer
Defaults index name to repo name (can override with -i)
Supports --ref to override branch/tag

3. Direct URL Syntax (`src/bin/index.ts`)

Pre-parse argument rewriting allows ctxc index <url> without the url subcommand
Transparently inserts url when a URL is detected

Backward Compatibility

✅ All existing subcommands work unchanged:

ctxc index github --owner augmentcode --repo context-connectors  # Still works

Testing

182 tests pass (19 new for URL parser)
Manually verified URL detection and indexing

Examples

# GitHub
ctxc index https://github.com/augmentcode/context-connectors
ctxc index https://github.com/augmentcode/context-connectors/tree/develop

# GitLab (including self-hosted)
ctxc index https://gitlab.com/group/project
ctxc index https://gitlab.mycompany.com/team/project

# Bitbucket
ctxc index https://bitbucket.org/workspace/repo

# Website
ctxc index https://docs.python.org/3/

# With options
ctxc index https://github.com/owner/repo -i custom-name --ref v2.0

Pull Request opened by Augment Code with guidance from the PR author

- Parse GitHub URLs (owner/repo, tree/branch, commit/sha) - Parse GitLab URLs (project path, subgroups, /-/tree/branch) - Parse Bitbucket URLs (workspace/repo, src/branch, branch/name) - Fallback to website source for unknown URLs - Extract default index names from URLs - Support self-hosted GitLab and Bitbucket instances - Export parseSourceUrl and ParsedUrl from @augmentcode/context-connectors/core - Add comprehensive unit tests (19 test cases) Agent-Id: agent-8394bd07-7a81-41d0-ac95-1ca62623e6fb

Adds command that auto-detects source type (GitHub, GitLab, Bitbucket, or website) from the URL and creates the appropriate source. Features: - Parses URL using parseSourceUrl() to determine source type - Supports --ref option to override URL-detected branch/tag - Supports -i/--index option to override default index name - Supports --store and --store-path options - Default index name derived from repo/project name - Graceful error handling for invalid URLs Agent-Id: agent-c9423996-94bb-4ab3-8311-ca0cc822da14

Adds pre-parse argument rewriting that auto-detects when a URL is passed directly to 'ctxc index' and transparently inserts the 'url' subcommand. Before: ctxc index url https://github.com/owner/repo After: ctxc index https://github.com/owner/repo Both syntaxes now work. Existing subcommands (github, gitlab, etc.) are unchanged and continue to work. Agent-Id: agent-ce81a04d-72f2-4289-8eb7-c3074d7d8030

richhankins · 2026-01-31T00:56:49Z

augment review

augment-app-staging · 2026-01-31T00:59:57Z

🤖 Augment PR Summary

Summary: This PR simplifies the indexing CLI by allowing users to pass a repository/website URL directly and having the tool auto-detect the source type.

Changes:

Added parseSourceUrl() in src/core/url-parser.ts to classify GitHub/GitLab/Bitbucket vs website URLs and extract source configs (including URL-derived refs).
Introduced ctxc index url <url> which instantiates the appropriate Source based on the parsed URL, with --ref override support.
Added argv rewriting so ctxc index <url> transparently maps to ctxc index url <url>.
Re-exported the parser + types from src/core/index.ts.
Added a Vitest suite covering common URL patterns across the supported providers.

Technical Notes: Self-hosted GitLab/Bitbucket URLs are supported via baseUrl, and URL-derived refs (e.g., /tree/<ref>) can be overridden by CLI flags.

_{🤖 Was this summary useful? React with 👍 or 👎}

augment-app-staging

Review completed. 3 suggestions posted.

Comment augment review to trigger a new review at any time.

augment-app-staging · 2026-01-31T00:59:58Z

src/core/url-parser.ts

+  }
+
+  const owner = pathParts[0];
+  const repo = pathParts[1];


URLs pasted from clone endpoints often end with .git; currently repo/projectName/defaultIndexName will include the suffix, which likely breaks GitHub/GitLab/Bitbucket API paths and yields odd default index keys. Consider stripping a trailing .git from repo/project segments when parsing.

Other Locations

src/core/url-parser.ts:125

src/core/url-parser.ts:152

_{🤖 Was this useful? React with 👍 or 👎, or 🚀 if it prevented an incident/outage.}

augment-app-staging · 2026-01-31T00:59:58Z

src/core/url-parser.ts

+  }
+
+  // GitLab (gitlab.com or hostname contains "gitlab")
+  if (hostname === "gitlab.com" || hostname.includes("gitlab")) {


Using hostname.includes("gitlab") / hostname.includes("bitbucket") can misclassify unrelated websites whose domain happens to contain those substrings and then fail parsing instead of falling back to website. Consider a more conservative self-hosted detection strategy to reduce false positives.

_{🤖 Was this useful? React with 👍 or 👎, or 🚀 if it prevented an incident/outage.}

augment-app-staging · 2026-01-31T00:59:58Z

src/bin/index.ts

-program.parse();
+// Auto-detect URL mode: ctxc index <url> -> ctxc index url <url>
+// This allows users to skip the 'url' subcommand when providing a URL directly
+const indexIdx = process.argv.indexOf("index");


The URL auto-rewrite only triggers when the URL is the first argument after index, so ctxc index -i myidx https://… won’t be rewritten and will likely error. Is that limitation intentional, or should the rewrite scan forward for the first non-option arg?

_{🤖 Was this useful? React with 👍 or 👎, or 🚀 if it prevented an incident/outage.}

richhankins added 3 commits January 30, 2026 00:20

augment-app-staging bot reviewed Jan 31, 2026

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Simplify CLI: Accept URLs directly for indexing#12

Simplify CLI: Accept URLs directly for indexing#12
richhankins wants to merge 3 commits intomainfrom
simplify-cli-arguments

richhankins commented Jan 30, 2026

Uh oh!

richhankins commented Jan 31, 2026

Uh oh!

augment-app-staging bot commented Jan 31, 2026

Uh oh!

augment-app-staging bot left a comment

Uh oh!

augment-app-staging bot Jan 31, 2026

Uh oh!

augment-app-staging bot Jan 31, 2026

Uh oh!

augment-app-staging bot Jan 31, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

Conversation

richhankins commented Jan 30, 2026

Summary

Before

After

Changes

1. URL Parser Module (src/core/url-parser.ts)

2. CLI URL Mode (src/bin/cmd-index.ts)

3. Direct URL Syntax (src/bin/index.ts)

Backward Compatibility

Testing

Examples

Uh oh!

richhankins commented Jan 31, 2026

Uh oh!

augment-app-staging bot commented Jan 31, 2026

Uh oh!

augment-app-staging bot left a comment

Choose a reason for hiding this comment

Uh oh!

augment-app-staging bot Jan 31, 2026

Choose a reason for hiding this comment

Uh oh!

augment-app-staging bot Jan 31, 2026

Choose a reason for hiding this comment

Uh oh!

augment-app-staging bot Jan 31, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

1. URL Parser Module (`src/core/url-parser.ts`)

2. CLI URL Mode (`src/bin/cmd-index.ts`)

3. Direct URL Syntax (`src/bin/index.ts`)