| name | tinyfish-search |
| description | Use when a chatbot request needs current web search or clean page fetching through TinyFish. This skill uses the local tinyfish CLI via the existing bash tool, reusing credentials from `tinyfish auth login`. |
TinyFish Search
Use the existing bash tool to call the local TinyFish CLI. Do not ask for or print API keys.
Check Auth
If TinyFish has not been used in the current session, first check:
tinyfish auth status
If it reports unauthenticated, tell the user to run tinyfish auth login in their shell.
If the tinyfish command is missing, install the CLI with:
npm install -g @tiny-fish/cli
Search
Run:
tinyfish search query "<query>"
Optional hints:
tinyfish search query "<query>" --location "<location>" --language "<language>"
The command returns JSON with query, results, and total_results. Each result includes fields such as position, title, url, site_name, and snippet.
When the user asks for a result count, trim the JSON result list yourself after the command returns.
Fetch Content
Run:
tinyfish fetch content get "<url>" --format markdown
For multiple URLs:
tinyfish fetch content get "<url1>" "<url2>" --format markdown
Use --format json when the user wants structured extraction, --links to include extracted links, and --image-links to include image links.
Response Rules
- Cite result URLs when answering from search or fetched pages.
- Keep the response to the fields the user requested.
- If the CLI fails, report the TinyFish error concisely and do not retry blindly.
