dsh-community-listening

Collecting public comments from scattered social platforms is one of the most time-consuming parts of community research. dsh-community-listening consolidates that workflow into a handful of DSH tool calls: discover content, gather publicly readable comments, persist them as JSONL corpora keyed by topic, run sentiment analysis and SimHash deduplication, cluster themes, and export a clean Markdown report.

The plugin ships as a standard DSH bundle that injects only the official tools service, so it runs identically in regular DSH profiles and DSH Desktop profiles with no Electron or desktop-internal dependencies. Each platform adapter maintains its own pacing, cooldown, and channel-health state; when a primary channel fails the plugin degrades gracefully and flags the actual backend and degraded status in every result. The analysis layer offers dictionary-rule and optional LLM sentiment scoring, embedding or keyword topic clustering, and SimHash near-duplicate detection, covering the full range from coarse filtering to detailed reading.

It is well suited for social-media researchers, product and operations teams, public-opinion analysts, and any DSH user who needs quantitative signals from publicly available comments. The plugin touches only content that is publicly readable and reuses the user's own browser session; it does not bypass CAPTCHAs, scrape behind login walls, or spoof device identity, keeping the compliance boundary clear and the setup straightforward.