coulomb/rein-openweights
4
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Record live proof-of-concept run via glas-harness's gateway

Browse source 
Real ext.bwrap sandbox, real OpenRouter call (qwen/qwen-2.5-72b-instruct),
real 2-turn tool-calling loop, real verified commit. Credential reused
the OPENROUTER_API_KEY already in the environment via credentials.py's
env-var short-circuit -- no OpenBao round trip needed this time.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
This commit is contained in:
tegwick 2026-07-26 13:46:24 +02:00
parent 4ee612140b
commit 4a82650ba5
2 changed files with 28 additions and 11 deletions
Download patch file Download diff file Expand all files Collapse all files

20
INTENT.md
View file

@ -50,9 +50,17 @@ owns the open-weight-model-specific agentic loop and tool execution.
Minimal agentic loop implemented and unit-tested (26 tests, all mocked Minimal agentic loop implemented and unit-tested (26 tests, all mocked
at the network/subprocess boundary): tool surface, OpenRouter client, at the network/subprocess boundary): tool surface, OpenRouter client,
credential acquisition, commit-verified success criterion, State Hub credential acquisition, commit-verified success criterion, State Hub
reporting. See `workplans/REIN-OW-WP-0001-bootstrap.md`. Not yet reporting. See `workplans/REIN-OW-WP-0001-bootstrap.md`.
executed against a real OpenRouter model or a real OpenBao instance —
those are human-triggered follow-ups (real API cost, real credentials). **Live-proven (2026-07-26):** run for real through glas-harness's
No glas-harness adapter (`reins/rein_openweights.py`, mirroring gateway (`glas_harness.reins.rein_openweights.ReinOpenWeights`) inside a
`reins/rein_aharness.py`) yet — that's a `GLAS-WP-0001` follow-up once real `ext.bwrap` sandbox, against the real OpenRouter API
this CLI is considered stable enough to shell out to. (`qwen/qwen-2.5-72b-instruct`) — a 2-turn tool-calling loop produced a
real, verified git commit. Credential reused the `OPENROUTER_API_KEY`
already present in the environment (the same one llm-connect resolves),
via this repo's own `credentials.py` env-var short-circuit — no OpenBao
round trip needed for this run.
The glas-harness adapter (`glas_harness/reins/rein_openweights.py`,
mirroring `reins/rein_aharness.py`) now exists and is what performed
this run.

19
workplans/REIN-OW-WP-0001-bootstrap.md
View file

@ -1,14 +1,23 @@
--- ---
id: REIN-OW-WP-0001 id: REIN-OW-WP-0001
title: "Bootstrap: minimal open-weight agentic loop" title: "Bootstrap: minimal open-weight agentic loop"
status: proposed status: active
--- ---
First working milestone for rein-openweights: prove an agentic tool-use First working milestone for rein-openweights: prove an agentic tool-use
loop against a current open-weight model via OpenRouter/llm-connect, loop against a current open-weight model via OpenRouter, comparable in
comparable in scope to rein-aharness's `green-commit-only` profile, ready scope to rein-aharness's `green-commit-only` profile, implementing the
to implement the glas-harness rein contract once glas-harness rein contract (`glas-harness/GLAS-WP-0001-T01`).
`glas-harness/GLAS-WP-0001-T01` defines it.
**Live-proven end-to-end (2026-07-26):** via glas-harness's gateway
(`glas_harness.reins.rein_openweights.ReinOpenWeights`), inside a real
`ext.bwrap` sandbox, against the real OpenRouter API with model
`qwen/qwen-2.5-72b-instruct` — a 2-turn tool-calling loop produced a
real git commit (`2effe57a`), verified via `git rev-parse HEAD`,
2449 tokens spent. Not a mock; real credentials, real API cost, run
with the user's explicit go-ahead. `executor-sandbox` (the disposable
target repo) was left untouched — bwrap only mutates its ephemeral
copy.
## Task: Pick the starter open-weight model + OpenRouter path ## Task: Pick the starter open-weight model + OpenRouter path