Skip to content

feat: add built-in fetch skill (read web page text) - #132

Draft
hallerite wants to merge 1 commit into
mainfrom
feat/fetch-skill
Draft

feat: add built-in fetch skill (read web page text)#132
hallerite wants to merge 1 commit into
mainfrom
feat/fetch-skill

Conversation

@hallerite

@hallerite hallerite commented Aug 16, 2026

Copy link
Copy Markdown
Member

Adds a fetch built-in skill: await fetch(url=...) returns readable page text (scripts/styles stripped, tags collapsed, output capped at max_chars), or an Error: ... string. Registered in _BUILTIN_SKILLS so RLM_SKILLS='...,fetch' enables it like edit/search.

Pairs with search (title/URL/snippet only): search finds URLs, fetch reads one. Motivated by large-scale eval: a search vs search+fetch comparison on bring-your-own-search tasksets needs a first-class fetch tool instead of emergent httpx. Tests: enable-stub, html→text stripping, non-http guard; suite 115 passed.

🤖 Generated with Claude Code

Pairs with search (title/URL/snippet only): search finds URLs, fetch reads one.
Async run(url, max_chars=8000) -> readable text (scripts/styles stripped, tags
collapsed, output capped), or an 'Error: ...' string on bad scheme / network error.
Registered in _BUILTIN_SKILLS so RLM_SKILLS='...,fetch' enables it like edit/search.

For the INTELLECT-5 search-tool study: openseeker/redsearcher are bring-your-own-search
tasksets, so the search-vs-search+fetch ablation needs a first-class fetch tool rather
than emergent httpx.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant