rt/wrolpi

mirror of https://github.com/lrnselfreliance/wrolpi.git synced 2026-10-26 10:51:36 -04:00

History

Roland c7bdedea09 ai: the /api/chat agent loop — SSE streaming, modes, tool execution Tool definitions are generated in-process from the /api/ai blueprint's own OpenAPI operations (operationId = tool name, description = tool description, dataclass bodies + declared parameters = one flat JSON schema) — the spec IS the catalog, nothing to drift. Manage endpoints never become tools. A mode is a server-enforced system prompt + tool allowlist (Help / Research / System per ai.md); the executor can only reach allowlisted /api/ai endpoints, path parameters are encoded and traversal-rejected so a crafted argument cannot escape an endpoint. POST /api/chat streams SSE (status/token/tool_call/tool_result/done/ error), starts llama-server on demand, validates tool args server-side with instructive errors the model can retry on, caps tool results with a paging hint, and bounds the loop by ai.yaml max_tool_calls plus a hard timeout. Caddy passes the stream unbuffered (flush_interval -1) in both deployments. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>	2026-09-24 18:23:21 -06:00
..
raspberrypios	ai: the /api/chat agent loop — SSE streaming, modes, tool execution	2026-09-24 18:23:21 -06:00
README.md	Combining install scripts into a new repair script. Removing create_db script in favor of repair script.	2023-10-03 12:15:29 -06:00

ai: the /api/chat agent loop — SSE streaming, modes, tool execution

Tool definitions are generated in-process from the /api/ai blueprint's
own OpenAPI operations (operationId = tool name, description = tool
description, dataclass bodies + declared parameters = one flat JSON
schema) — the spec IS the catalog, nothing to drift.  Manage endpoints
never become tools.  A mode is a server-enforced system prompt + tool
allowlist (Help / Research / System per ai.md); the executor can only
reach allowlisted /api/ai endpoints, path parameters are encoded and
traversal-rejected so a crafted argument cannot escape an endpoint.

POST /api/chat streams SSE (status/token/tool_call/tool_result/done/
error), starts llama-server on demand, validates tool args server-side
with instructive errors the model can retry on, caps tool results with
a paging hint, and bounds the loop by ai.yaml max_tool_calls plus a
hard timeout.  Caddy passes the stream unbuffered (flush_interval -1)
in both deployments.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

2026-09-24 18:23:21 -06:00

raspberrypios

ai: the /api/chat agent loop — SSE streaming, modes, tool execution

2026-09-24 18:23:21 -06:00

README.md

Combining install scripts into a new repair script. Removing create_db script in favor of repair script.

2023-10-03 12:15:29 -06:00

README.md

Configuration Files

The files in etc are files necessary to build an on-the-metal WROLPi. Raspberry Pi has more config files because it is assumed to be the main OS. Debian uses the Raspberry Pi configs when possible.