[Kralizek]

Summary

Adds the generic Hooksmith capabilities needed to compose HTML metadata enrichment without coupling provider integrations to page scraping.

@hooksmith/http

• adds fetchText(...) for arbitrary-method text responses
• adds getText(...) as the GET convenience transformer
• mirrors the existing fetchJson/getJson transformer conventions
• supports headers, arbitrary request bodies, mapping, and the existing non-2xx failure behavior
• adds focused tests for GET, arbitrary methods/body mapping, and unsuccessful responses

New @hooksmith/html

Introduces the newly registered @hooksmith/html package for HTML parsing and metadata extraction utilities.

Initial surface:

• OpenGraphMetadata
• parseOpenGraph() — pure pipeline transformer from HTML text to metadata
• getOpenGraph() — convenience transformer that fetches HTML through @hooksmith/http and parses the result

The initial Open Graph parser extracts:

• og:title
• og:description
• og:image
• og:url
• og:type
• og:site_name

It also falls back to <title> and meta[name=description], decodes common HTML entities, and resolves relative Open Graph image/URL values when a base URL is known.

Design boundary

The intended layering is:

@hooksmith/http  -> transport / text retrieval
@hooksmith/html  -> HTML interpretation / metadata extraction
provider package -> provider-specific serialization

This deliberately keeps Open Graph discovery out of social/provider packages such as @hooksmith/bluesky.

Compatibility

Existing HTTP APIs are unchanged. The new text transformers and HTML package are additive.

[Copilot]

Copilot review overview

🟡 Changes recommended

Unresolved HTML parsing and redirect URL handling issues need correction before approval.

Review effort: Lite
Findings: 2 · 2

Open (4)

• Handle > characters inside quoted meta attributes · New
• Decode HTML entities in a single pass · New
• Rename misleading getOpenGraph example variable · New
• Document public fetchText and getText APIs · New

What changed in this PR

Adds text HTTP transformers and a new HTML package for Open Graph metadata extraction.

Changes:

• Adds fetchText and getText with mapping, headers, bodies, and error handling.
• Introduces Open Graph parsing and fetching utilities.
• Registers the new workspace package with tests and documentation.

File	Description
extensions/​http/​types.ts	Defines text transformer types.
extensions/​http/​transformers.ts	Implements text HTTP transformers.
extensions/​http/​mod.ts	Exports the new HTTP APIs.
extensions/​http/​fetch_text_test.ts	Tests text transformer behavior.
extensions/​html/​README.md	Documents the HTML package.
extensions/​html/​mod.ts	Implements Open Graph parsing and fetching.
extensions/​html/​mod_test.ts	Tests metadata extraction.
extensions/​html/​deno.json	Defines the HTML package.
deno.json	Registers the HTML workspace package.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[Copilot]

Copilot review overview

🔵 Needs a closer look

Three unresolved moderate correctness issues remain in the HTML metadata parser.

Review effort: Lite
Findings: None

Resolved since last review (4)

• Decode HTML entities in a single pass
• Handle > characters inside quoted meta attributes
• Document public fetchText and getText APIs
• Rename misleading getOpenGraph example variable

Previously missed (1)

In code that hasn't changed since last review

Avoid extracting metadata from comments and script content

extensions/​html/​mod.ts:80

Because this regex scans raw source rather than tokenizing HTML, it also matches <meta> text inside comments or script/string content. For example, a commented-out og:title before the real tag is stored first and wins, so pages with embedded examples can return incorrect metadata. Ignore comments/script blocks or use a DOM-aware parser before extracting tags.