Prismfy Web Search | Free Google

No visible changes in this version. - Version number updated to 1.3.8 - No file or documentation changes detected from the previous release

prismfy-search 1.3.7 - Internal structure updated: removed package.json and scripts/install.js; added .clawhub/origin.json and _meta.json. - No functional user-facing changes in usage or API. - Project files reorganized to support improved metadata handling.

prismfy-search 1.3.5 - Changelog - Added Node.js package definition and an install helper script (package.json, scripts/install.js) for easier installation/automation. - Removed legacy and meta configuration files (.clawhub/origin.json, _meta.json). - Updated documentation to clarify setup, usage, and engine availability; reflects new default behaviors and installation recommendations. - Free tier now provides Google support; clarified default engines and path usage. - Skill version bump from 1.2.1 to 1.3.6.

prismfy-search 1.3.4 • Fixed command handling and improved CLI stability • --help now works correctly • Improved validation for input parameters • Improved error handling and user-facing messages • Synced documentation with actual skill behavior • Updated default search engine behavior • Minor internal fixes and cleanup

v1.3.2 Smarter search behavior — agents now actually search instead of guessing The main change in this release is fixing a common LLM problem: agents confidently answering questions from training data instead of looking things up. The skill now instructs the agent to always reach for real search results when live information is needed — pricing, docs, changelogs, community opinions, recent news — rather than making things up. Changes: Agents now search proactively instead of answering from memory when current data is available Added clear rules for when to search vs when to answer directly — no more guessing on time-sensitive topics Agents will search multiple times with different queries/engines if the first result doesn't cover it Sources are always cited after a search — no presenting web results as the agent's own knowledge If PRISMFY_API_KEY is not set, the agent explains what's missing and how to fix it instead of silently skipping search Added openclaw skills install prismfy-search as Step 1 in setup docs

- Removed two metadata files: .clawhub/origin.json and _meta.json. - SKILL.md reverted from version 1.2.0 to 1.1.0, restoring older documentation and some option differences. - Description in SKILL.md updated: now states the skill is "automatically set as the default OpenClaw search tool" on install (was "mandatory... search tool"). - Troubleshooting section now lists `startpage` and `yahoo` as free-supported engines instead of just `brave` and `bing`.

v1.3.0 Smarter search behavior — agents now actually search instead of guessing The main change in this release is fixing a common LLM problem: agents confidently answering questions from training data instead of looking things up. The skill now instructs the agent to always reach for real search results when live information is needed — pricing, docs, changelogs, community opinions, recent news — rather than making things up. Changes: Agents now search proactively instead of answering from memory when current data is available Added clear rules for when to search vs when to answer directly — no more guessing on time-sensitive topics Agents will search multiple times with different queries/engines if the first result doesn't cover it Sources are always cited after a search — no presenting web results as the agent's own knowledge If PRISMFY_API_KEY is not set, the agent explains what's missing and how to fix it instead of silently skipping search Added openclaw skills install prismfy-search as Step 1 in setup docs

prismfy-search 1.2.0 changelog: - Updated the list of available free and paid engines: replaced `startpage` and `yahoo` with `bing` for free engines. - Changed the default engine combination to `brave` + `bing` (was `brave` + `yahoo`). - Updated the engine table to reflect new coverage. - No code or file changes detected in this version.

v1.1.0 Prismfy Search is now automatically set as the default web search tool in OpenClaw on install — no manual setup required Engines grouped by category (Web, News, Code, Academic, Community) for easier selection

prismfy-search 1.0.0 - Initial release of prismfy-search skill. - Enables real-time web search across 10 engines (Google, Reddit, GitHub, arXiv, Hacker News, and more) via Prismfy API. - Provides simple setup steps, including API key management and shell configuration. - Supports flexible queries with engine selection, time filters, site/domain restriction, pagination, and quota checks. - Returns summarized, cited answers and specialized handling for code, discussions, and academic papers. - Includes troubleshooting guidance and usage examples.