Activity

1. senamakel

   assigned

   oxoxDev

   on Nov 9, 2026
2. tinysweeper

   added

   priority: p3Whenever. Cosmetic, a nicety, or a cleanup with no user visible effect.Whenever. Cosmetic, a nicety, or a cleanup with no user visible effect.

   on Nov 9, 2026
3. tinysweeper commented on Nov 9, 2026

   

   tinysweeperBot

   on Nov 9, 2026

   More actions

   Tracking issue for a one-off triage pass over every open PR and issue (Nov 9-13 milestone). It is a process/coordination task rather than a bug or feature request, so it needs no code fix; low priority for maintainers.

   Labelled priority: p3.
4. tinysweeper

   added the

   Task

   issue type on Nov 9, 2026
5. M3gA-Mind

   assigned

   M3gA-Mind

   and unassigned

   oxoxDev

   on Nov 11, 2026
6. M3gA-Mind commented on Nov 11, 2026

   

   M3gA-Mind

   on Nov 11, 2026

   Collaborator

   More actions

   Status update, 2026-11-11 14:50 IST: merged #599. The tables below are the triage as it was done; rows for those items are now settled.

   tinyagents open-PR triage (team#45 / tinyagents#597), 2026-11-11

   Official main b6ed460 (CI green), which is also openhuman main's pin. All PRs were test-merged onto it; the mergeable ones were built and tested on our runner.

   PR	Verdict	Reason	Next step
   #599 install-action 2.88.1→2.90.2 (dependabot)	ready to merge	Pinned-SHA bump in the Coverage job; 0 behind main, all checks green	Maintainer merges
   #571 retain classified stop cause (draft)	ready to merge once un-drafted	Correct upstream fix for what openhuman #8146 works around on the host; on main it passes 3311/0, including 3 new tests	Author un-drafts; maintainer reviews and merges; then openhuman can drop its host copy
   #588 deliver-first turn after a guard halt	needs changes	Breaks main's a_later_registered_after_tool_sees_the_marker_on_refused_calls (a second "halted"); no own test; 2 Major CodeRabbit threads unanswered	Author fixes the contract, adds a test and answers the threads; maintainer weighs it against openhuman's post-loop close
   #567 queue edits + durable run inbox	needs changes	No longer compiles on main (run_queue now uses Queued<T>; E0308/E0277 at run_queue/mod.rs:477-523); no openhuman consumer	Author rebases and adapts; maintainer decides on the new API
   #583 structural tools + durable input (draft)	needs changes	Bundles #567 and #571; pins a tinytools commit off tinytools main; conflicts; no CI run	Author splits it, upstreams the parser to tinytools, re-cuts a tools-only PR
   #517 tinyinference bump for OCI	needs changes (blocked)	Pins the head of tinyinference#122, which is still open and diverged from tinyinference main; conflicts in the gitlink and Cargo.lock	After #122 merges, re-cut on main with a main-commit gitlink

   Issues

   Checked by reading the code at main b6ed460 (= openhuman's pin) and openhuman e63ef876b6; no runs.

   Issue	Verdict	Reason	Next step	Priority
   #598 hosted errors replaced by a fixed string	still real	hosted_error_message still returns fixed strings (runtime/agent.rs:198-206); HostedError has no provider class; From drops run	Carry the provider class on HostedError, then map it in openhuman's classifier; add a 401 test	p2
   #394 tier-aware pricing projections	still real	estimate_cost_usd tiers M3 ≤512K, but flat_catalog_entry projects flat 2× rates with no tiers; openhuman uses both paths	Fill tiers for M3; audit the Gemini 200K and GPT 272K thresholds	p3
   #393 cached-input rate audit	still real	Every named row is still 0.25× input under one PRICING_AS_OF = "2026-07"; the repo's models.dev snapshot says 0.1× for Gemini 2.5	Verify each row against provider docs; add a source and date per row	p2
   #322 agent_loop behaviours from the #318 review	still real (partly)	(c) fixed by #339, (d) fixed by #532; (a), (b) and (e) still present; (f) ambiguous	Split (a), (b), (e) and (f) into separate issues; (b) is the clearest bug	p3