feat: probe-band-links v3 — engine-agnostic web-search fallback#629
Merged
Conversation
BreakableHoodie
added
enhancement
BreakableHoodie
added
enhancement
|
You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard. |
Bands unresolved by canonical-URL guessing now get one real web search: DuckDuckGo's keyless HTML endpoint by default (zero setup), upgraded to Google Programmable Search when GOOGLE_CSE_KEY/ID are set. Search hits land in REVIEW only — a result link can't prove identity the way a bandcamp location can. Acid-tested on the exact case the built-in search tool missed (Blue Freezie: DDG's top hit is the band's real page). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
BreakableHoodie
force-pushed
the
feat/probe-links-google-cse
branch
from
July 17, 2026 19:43
31646e3 to
a9f085e
Compare
![github-advanced-security[bot]](https://avatars.githubusercontent.com/in/57789?s=60&v=4){kind=small}
js/incomplete-url-substring-sanitization (high): filtering result
URLs with `!u.includes("duckduckgo.com")` is defeated by hosts like
`duckduckgo.com.evil.test` or `evil.test/?x=duckduckgo.com`. Parse
each URL and compare the real hostname (exact or `.duckduckgo.com`
suffix); unparseable URLs are dropped.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Follow-up to the link-probe tooling merged in #627, per Dre's feedback that manual web searches find pages the probe missed. Bands unresolved by canonical bandcamp URL-guessing now get one real web search — engine-agnostic: DuckDuckGo's keyless HTML endpoint by default (works with zero setup), auto-upgrading to Google Programmable Search when
GOOGLE_CSE_KEY/GOOGLE_CSE_IDenv vars exist (100 free queries/day). Search hits are bucketed REVIEW only, never AUTO — a result link can't prove band identity the way an exact name match + Ontario location on a bandcamp page can.Live validation this session: DDG's top hit for
"Blue Freezie" band bandcampis the band's real page (the exact case the built-in search tool failed on); Vox Confidential resolved AUTO via URL-guess + full-body location extraction ("Kitchener, Ontario") — the doctrine-safe path working end-to-end.What changed
scripts/probe-band-links.mjsddgSearch()(keyless, music-domain-ranked, deduped, top 5) +googleSearch()(CSE, optional); unresolved bands route through the active engine; REVIEW rendering shows per-engine hits; throttled like the bandcamp probesSecurity / correctness notes
Read-only outbound HTTP from a local script; no credentials required or stored (CSE keys read from env only, never written). No app code touched.
Verification
validate:openapi: n/aBuilt by Theo · Reviewed by Theo · 🤖 Claude Code