- Home
- Skills
- Anthropics
- Knowledge Work Plugins
- Source Management
source-management_skill
- Python
- Official
7.4k
GitHub Stars
1
Bundled Files
2 months ago
Catalog Refreshed
4 months ago
First Indexed
Readme & install
Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.
Installation
Preview and clipboard use veilstrat where the catalogue uses aiagentskills.
npx veilstrat add skill anthropics/knowledge-work-plugins --skill source-management- SKILL.md5.6 KB
Overview
This skill manages connected MCP sources for enterprise search, detects which sources are available, and helps users connect new sources. It controls source ordering by query intent and maintains rate-limiting and health awareness so search results are reliable and transparent.
How this skill works
It inspects available MCP tool prefixes to discover connected sources and maps each source to its key capabilities (chat, email, cloud storage, project tracker, CRM, knowledge base). It applies priority rules based on query type to weight results, skips or deprioritizes rate-limited sources, and reports which sources were searched. It also guides users through connecting new MCP sources and updating configuration.
When to use it
- When search results seem incomplete or you suspect missing sources
- When onboarding a new data source or verifying current connections
- When tailoring search behavior for a specific query type (decisions, status, documents, people, policies)
- When handling slow or failing source responses to avoid blocking searches
- When preparing digests and you need source coverage and time-range awareness
Best practices
- Prefer targeted queries to avoid broad scans that trigger rate limits
- Weight sources by query intent instead of excluding them entirely
- Respect rate-limit responses (do not retry immediately) and continue with other sources
- Include which sources were searched when presenting results so users understand scope
- Cache recent query results to reduce redundant API calls
Example use cases
- User runs a company-wide search and the skill shows which sources are included and suggests connecting missing ones
- A decision-related query prioritizes chat and email so meeting outcomes surface first
- During a digest generation, the skill detects a rate-limited source and notes which time ranges were scanned before the limit
- Admin adds a new MCP server; the skill detects the new source and includes it in subsequent searches automatically
- User asks who owns a project; the skill orders results to surface chat authors and task assignees first
FAQ
The skill lists connected MCP sources by detecting available tool prefixes and shows their current status (available, not connected, rate limited).
What happens if a source is rate limited?
The skill stops immediate retries, continues querying other sources, caches progress for digests, and informs the user which source was rate limited and which results are shown.