The existing MCP feature doc covers the technical mechanics of agents connecting to a JSS pod, including a short "How Charlie works" section about the bot layout. What's missing is a standalone doc for the Charlie pattern itself:
- The TBL 2017 vision and Inrupt 2024 prototype lineage
- What the personal-assistant pattern looks like running on standards-grounded substrate today
- The verbatim agent-meets-WAC interaction trace (naïve probe → examine ACL → comprehend scope)
- The architectural thesis: consent as a language the agent reads, not a fence it runs into
- A starting
SKILL.md template
- Forward roadmap (scoped credentials, agent-to-agent flows, federation)
This complements rather than duplicates the MCP doc , MCP is the how, Charlie is the what and why. Adds a layered docs surface: MCP for connection mechanics, Charlie for the personal-assistant pattern, with explicit cross-links between them.
The existing MCP feature doc covers the technical mechanics of agents connecting to a JSS pod, including a short "How Charlie works" section about the bot layout. What's missing is a standalone doc for the Charlie pattern itself:
SKILL.mdtemplateThis complements rather than duplicates the MCP doc , MCP is the how, Charlie is the what and why. Adds a layered docs surface: MCP for connection mechanics, Charlie for the personal-assistant pattern, with explicit cross-links between them.