How long does the integration take?

Change one base URL in an OpenAI-compatible client, or install the Python or TypeScript SDK; the quickstart routes a first request in a few minutes.

Will this lock us into another vendor?

No. The router is Apache-2.0 open source, your provider keys stay yours, and the API is OpenAI-compatible, so leaving is a change of base URL.

What happens when a provider fails?

The router falls back to the next model your policy allows and records the fallback in the trace, so an outage costs a retry, not an incident.

tiangong-kb-course-fulltext-fetch - Skill

Name: tiangong-kb-course-fulltext-fetch
Author: tiangong-ai

Imported from tiangong-ai/agent-skills (tiangong-kb-course-fulltext-fetch/SKILL.md). Install upstream with npx skills add tiangong-ai/agent-skills --skill tiangong-kb-course-fulltext-fetch. Copyright stays with the author.

Tiangong KB Course Fulltext Fetch

Use this skill when the task has a Tiangong KB course document_id and tags, and needs the full processed text for that document.

Boundary

Always call the Tiangong AI CLI:

npx @tiangong-ai/cli@0.0.19 kb course fulltext --document-id <document_id> --tags <tags>

Do not call S3, NAS, Supabase, Pinecone, or OpenSearch directly from the skill. The CLI owns the S3 path rule, list/get behavior, and AWS SDK credential handling.

Prerequisites

The wrapper defaults to npx @tiangong-ai/cli@0.0.19; users do not need a preinstalled CLI. Set TIANGONG_AI_CLI or TIANGONG_AI_CLI_BIN only to override the CLI entrypoint.
AWS credentials must already be available in the environment through the AWS SDK default chain, such as AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY, AWS_SESSION_TOKEN, AWS_PROFILE, AWS_REGION, or AWS_DEFAULT_REGION.
If credentials live in a dotenv file, pass env_file. Loaded dotenv values only fill unset environment variables; explicit JSON fields such as bucket, prefix, and region are passed as CLI flags and take precedence. Never print secret values.

Fetch Full Text

Pass the course document_id and tags:

./scripts/course_fulltext_fetch.sh '{
  "document_id": "000125ed-c4d9-4fe3-9383-281162406d66",
  "tags": "thu_humanities"
}'

The wrapper calls:

npx @tiangong-ai/cli@0.0.19 kb course fulltext --document-id <document_id> --tags <tags>

To save the full text:

./scripts/course_fulltext_fetch.sh '{
  "document_id": "000125ed-c4d9-4fe3-9383-281162406d66",
  "tags": "thu_humanities"
}' ./fulltext.txt

To return CLI metadata and text as JSON:

./scripts/course_fulltext_fetch.sh '{
  "document_id": "000125ed-c4d9-4fe3-9383-281162406d66",
  "tags": "thu_humanities",
  "json": true
}'

Input Fields

document_id or documentId: required course document id.
tags or tag: required course tag, such as thu_humanities.
output_file: optional output path. The second wrapper argument takes precedence when both are provided.
env_file: optional dotenv file to load before calling the CLI.
json: true to return the CLI JSON payload instead of plain text.
bucket, prefix, region: optional overrides for the CLI command.

When using the returned text in reader-facing prose, cite source titles or paths when useful, but do not describe internal S3 paths, AWS credentials, or retrieval mechanics unless the user asks for implementation details.

Install tiangong-kb-course-fulltext-fetch and it becomes one more candidate the router can pick - when it fits.

A listing is a routing target with a manifest: what it does, which domains it covers, what it costs and who publishes it. Once installed it sits beside your own models and tools, is scored like any other candidate for each request, and shows up in the trace when it wins. Ratings come from workspaces that installed it, one per account.

{ "ocm": "1", "id": "tiangong-ai-agent-skills-tiangong-kb-course-fulltext-fetch", "kind": "skill", "name": "tiangong-kb-course-fulltext-fetch", "description": "Fetch full text for a Tiangong KB course document through the Tiangong AI CLI. Use when a task has a course document_id and tags and needs the complete processed text.", "publisher": "tiangong-ai", "version": "1.0.0", "capabilities": { "domains": [ "math" ], "tags": [ "skill-md", "skills-sh" ], "languages": [ "en" ] }, "quality_prior": 0.6, "examples": [ "Fetch full text for a Tiangong KB course document through the Tiangong AI CLI. Use when a task has a course document_id and tags and needs the complete processed text." ], "primary": false, "metadata": { "source": { "provider": "skills.sh", "repository": "https://github.com/tiangong-ai/agent-skills", "path": "tiangong-kb-course-fulltext-fetch/SKILL.md", "ref": "HEAD", "url": "https://github.com/tiangong-ai/agent-skills/blob/HEAD/tiangong-kb-course-fulltext-fetch/SKILL.md", "key": "tiangong-ai/agent-skills/tiangong-kb-course-fulltext-fetch/SKILL.md" } }, "instructions": "# Tiangong KB Course Fulltext Fetch\n\nUse this skill when the task has a Tiangong KB course `document_id` and `tags`,\nand needs the full processed text for that document.\n\n## Boundary\n\nAlways call the Tiangong AI CLI:\n\n```bash\nnpx @tiangong-ai/cli@0.0.19 kb course fulltext --document-id <document_id> --tags <tags>\n```\n\nDo not call S3, NAS, Supabase, Pinecone, or OpenSearch directly from the skill.\nThe CLI owns the S3 path rule, list/get behavior, and AWS SDK credential\nhandling.\n\n## Prerequisites\n\n- The wrapper defaults to `npx @tiangong-ai/cli@0.0.19`; users do not need a\n preinstalled CLI. S", "cost": { "context_tokens": 638 } }

tiangong-kb-course-fulltext-fetch

About

Tiangong KB Course Fulltext Fetch

Boundary

Prerequisites

Fetch Full Text

Input Fields

Install tiangong-kb-course-fulltext-fetch and it becomes one more candidate the router can pick - when it fits.

What it costs to adopt - in engineering time.

One base URL

No vendor lock-in

Failover built in

Use it

Manifest

Reviews