int_steering #112

Closed
rawan wants to merge 0 commits from int_steering into integration
Member

#93

#93
fix: tables markdown rendering
Some checks failed
Verify / verify (pull_request) Has been cancelled
Verify / verify (push) Has been cancelled
eca59dba67
Merge remote-tracking branch 'origin/development' into development_fix_markdown_table_rendering
Some checks failed
Verify / verify (push) Has been cancelled
Verify / verify (pull_request) Has been cancelled
d3531bae79
# Conflicts:
#	crates/hero_shrimp_web/static/assets/app.BWKElLkA.js
#	crates/hero_shrimp_web/static/assets/app.CaZse4bC.js
#	crates/hero_shrimp_web/static/assets/app.D1oK3cUT.js
#	crates/hero_shrimp_web/static/index.html
Merge remote-tracking branch 'origin/development' into development_fix_markdown_table_rendering
Some checks failed
Verify / verify (pull_request) Successful in 38m5s
Verify / verify (push) Failing after 2h5m18s
e3d135573c
# Conflicts:
#	crates/hero_shrimp_web/static/assets/app.BWKElLkA.js
#	crates/hero_shrimp_web/static/assets/app.C84XiDPf.js
#	crates/hero_shrimp_web/static/assets/app.ZKpvRLfm.js
#	crates/hero_shrimp_web/static/index.html
Reviewed-on: #82
ci: add integration branch to triggers alongside development
Some checks failed
Build Linux / build-linux (push) Successful in 10m17s
Verify / verify (push) Has been cancelled
2754adee87
The optimistic ask bubble was only cleared when a reply arrived, so
after refresh() re-fetched the persisted ask it rendered alongside the
optimistic copy during the in-flight window. Clear pendingAsk once the
real ask is present in fetched messages, and guard focusedThread() from
appending a duplicate ask while keeping the working indicator.

#72
feat(crew): group thread into ask→reply turn cards with timeline rail
All checks were successful
Verify / verify (push) Successful in 40m51s
Verify / verify (pull_request) Successful in 40m39s
171226b4ae
Render the focused agent thread as conversation turns instead of a flat
timestamp-sorted list. Each ask is paired with its reply (matched by
correlation_id, already exposed by crew.messages) inside one parent card,
joined by a centred left-rail connector: an accent dot for the ask and a
good dot for the reply, linked by a continuous vertical line. This stops
the thread ever showing reply-after-reply or ask-after-ask.

An ask still in flight with no reply yet gets a synthetic "working…" reply
(its dot pulses), so a live ask never renders as a lone bubble.

#72
Merge pull request 'fix(crew): prevent duplicate ASK bubble for new question' (#84) from development_fix_crew_duplicate_ask into integration
All checks were successful
Build Linux / build-linux (push) Successful in 10m22s
Verify / verify (push) Successful in 39m48s
fae1d20b9a
Reviewed-on: #84
fix: fixed socket mismatch, updated docs to only use lab
All checks were successful
Verify / verify (pull_request) Successful in 12m8s
Verify / verify (push) Successful in 30m28s
513ca6094f
Merge pull request 'fix: fixed socket mismatch, updated docs to only use lab' (#88) from int_fix_socket into integration
All checks were successful
Build Linux / build-linux (push) Successful in 4m20s
Verify / verify (push) Successful in 30m49s
659622bde5
Reviewed-on: #88
Merge remote-tracking branch 'origin/main' into integration
All checks were successful
Build Linux / build-linux (push) Successful in 10m4s
Verify / verify (push) Successful in 30m17s
0a3920bf18
fix: veridict box
Some checks failed
Verify / verify (push) Successful in 32m52s
Verify / verify (pull_request) Has been cancelled
dc21a00354
fix(web): stamp explicit incomplete/stop_reason instead of UI prose matching
All checks were successful
Verify / verify (pull_request) Successful in 12m47s
Verify / verify (push) Successful in 32m54s
f6c9572a2d
Replace the UI's brittle prose-matching (isIncompleteReply over ~8 hardcoded
phrases) with an explicit incomplete/stop_reason field stamped by the server on
every turn/job reply. The UI keys its Continue/Retry affordance off that field
and falls back to the prose heuristic only when the field is absent (older
daemon). Detection reuses existing server signals — the structured step-limit
event, failure_kind, and a single server-owned reply_indicates_incomplete_run
fallback — so there is no new prose matcher.

#40
bump rust to 1.96
Some checks failed
Verify / verify (push) Has been cancelled
Verify / verify (pull_request) Has been cancelled
e0bc63d320
Merge remote-tracking branch 'origin' into int_fix_veridict
All checks were successful
Verify / verify (pull_request) Successful in 12m31s
Verify / verify (push) Successful in 30m58s
86e60749ae
Merge pull request 'fix: veridict box' (#97) from int_fix_veridict into integration
All checks were successful
Build Linux / build-linux (push) Successful in 10m44s
Verify / verify (push) Successful in 31m11s
c4992eeee4
Reviewed-on: #97
Merge remote-tracking branch 'origin/integration' into development_structured_incomplete_field
All checks were successful
Verify / verify (push) Successful in 12m34s
Verify / verify (pull_request) Successful in 32m15s
0bfef5c1bb
# Conflicts:
#	crates/hero_shrimp_web/static/assets/app.C9ryv1Pi.js
#	crates/hero_shrimp_web/static/assets/app.CLFV4RoG.js
#	crates/hero_shrimp_web/static/assets/app.DsplZEWO.js
#	crates/hero_shrimp_web/static/index.html
Merge pull request 'fix(web): stamp explicit incomplete/stop_reason instead of UI prose matching' (#96) from development_structured_incomplete_field into integration
All checks were successful
Build Linux / build-linux (push) Successful in 11m12s
Verify / verify (push) Successful in 30m58s
432df99bdb
Reviewed-on: #96
Merge branch 'integration' of ssh://forge.ourworld.tf/lhumina_code/hero_shrimp into integration
Some checks are pending
Build Linux / build-linux (push) Successful in 4m28s
Verify / verify (push) Waiting to run
f6980accd4
fix: fix auto approve checkbox
All checks were successful
Verify / verify (push) Successful in 18m25s
7229461621
ci: canonical lab-publish workflow (build main/development/integration)
Some checks failed
Verify / verify (push) Has been cancelled
Build Linux / build-linux (push) Has been cancelled
lab publish / publish (push) Has been cancelled
c0a4301db3
Publishes musl-x86_64 binaries to per-branch releases (latest,
latest-dev, latest-integration) and installs lab from the matching
hero_skills branch (clone + build via --branch). Triggers only on push
to these three branches.
ci: trigger lab-publish run
Some checks failed
Build Linux / build-linux (push) Successful in 4m30s
lab publish / publish (push) Failing after 9m41s
Verify / verify (push) Successful in 11m53s
02f4ac677d
rebasing to integration
Reviewed-on: #103
Replaces lab-publish.yaml with a single lab-release workflow that pulls the
prebaked lab-builder image and publishes per-branch releases (main=stable,
development/integration=pre-release). No per-run toolchain/lab install.
fix: auto approve added, toggle the mode from yolo added
Some checks failed
Verify / verify (push) Failing after 23s
15936ab603
fix: format
All checks were successful
Verify / verify (push) Successful in 8m59s
Verify / verify (pull_request) Successful in 16m20s
5d408f9c0d
Merge pull request 'int_auto_approve' (#108) from int_auto_approve into integration
Some checks failed
Build Linux / build-linux (push) Successful in 9m23s
Verify / verify (push) Successful in 9m43s
lab release / release (push) Failing after 15m22s
a6f795b9b8
Reviewed-on: #108
ci: canonical lab-release (cargo check + multi-arch + hero.releaser)
Some checks failed
Build Linux / build-linux (push) Successful in 4m26s
lab release / release (push) Failing after 29m42s
Verify / verify (push) Successful in 11m25s
4dc1c739e5
Remove agent step/iteration limiting entirely and make the spend budget
the only execution limiter:

- resolve_iteration_budget always returns an unbounded ceiling, so the
  agent loop (and subagents) are never stopped by an iteration count
- max_iterations/hard_max_iterations and the duplicate per-job budget
  fields (per_job_usd, per_job_max_calls, autonomy.job_budget_usd) are
  deprecated no-ops, still parsed for backward compatibility
- job.start refuses a run when no budget is configured (with the
  HERO_SHRIMP_BUDGET_OVERRIDE escape hatch); default.yml ships
  budget.daily_usd so runs stay bounded out of the box
- the per-call cap and the job-start gate read the live global config so
  a budget edited from the UI takes effect without a restart

#98
#98
#98
#98
#98
fix(web): budget save dirty-guard, budget-stop shows raise-budget only
All checks were successful
Verify / verify (push) Successful in 35m29s
Verify / verify (pull_request) Successful in 35m49s
919b61724f
#98
Merge pull request 'feat(budget): make spend budget the sole run limiter' (#101) from development_unify_budget_limit into integration
Some checks failed
Verify / verify (push) Successful in 9m17s
Build Linux / build-linux (push) Successful in 11m53s
lab release / release (push) Failing after 1h6m17s
efcb9613aa
Reviewed-on: #101
config.get now exposes a models array (alias, tier, ctx, caps) from the
global model catalog. The Settings modal per-phase model datalist and
the Crew hire-modal Model field are populated from it, with the old
hardcoded list kept only as a fallback. Free-text entry stays allowed.

#105
fix(web): replace model datalists with a combo-box dropdown showing the full catalog
All checks were successful
Verify / verify (push) Successful in 13m19s
Verify / verify (pull_request) Successful in 29m55s
a9da576b1a
Merge pull request 'fix(web): populate model select boxes from the configured catalog' (#106) from integration_fix_model_select_boxes into integration
All checks were successful
Build Linux / build-linux (push) Successful in 5m29s
Verify / verify (push) Successful in 11m59s
lab release / release (push) Successful in 1h11m16s
d9ea8113fc
Reviewed-on: #106
fix: steering
All checks were successful
Verify / verify (push) Successful in 38m18s
Verify / verify (pull_request) Successful in 39m4s
15705bf9e7
rawan changed target branch from main to integration 2026-06-11 14:33:05 +00:00
rawan closed this pull request 2026-06-11 14:33:19 +00:00
All checks were successful
Verify / verify (push) Successful in 38m18s
Verify / verify (pull_request) Successful in 39m4s

Pull request closed

Sign in to join this conversation.
No reviewers
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
lhumina_code/hero_shrimp!112
No description provided.