This website requires JavaScript.
2a43210f38
Merge pull request 'feat(run): critic owns the deadline — MaxRuntime becomes the soft trigger' (#21 ) from feat/critic-owns-deadline into main
steve
2026-06-30 15:56:31 +00:00
79ce833dd7
fix(run): address round-2 gadfly nits (max(), drop dead soft fallback, decouple doc)
steve
2026-06-30 11:54:38 -04:00
cb4c612461
fix(run): address gadfly review of the critic-deadline PR
steve
2026-06-30 11:32:46 -04:00
5b5ee4148e
feat(run): critic owns the deadline — MaxRuntime becomes the soft trigger
steve
2026-06-30 11:03:40 -04:00
31f9078915
Merge pull request 'feat(run): durable checkpoint + resume (wire Ports.Checkpointer)' (#20 ) from feat/kernel-checkpoint into main
steve
2026-06-29 20:44:17 +00:00
38d656ec71
fix(run): address gadfly review of the checkpoint PR
steve
2026-06-29 16:34:42 -04:00
899059a791
feat(run): durable checkpoint + resume (wire Ports.Checkpointer)
steve
2026-06-29 16:04:06 -04:00
c071ed4996
Merge pull request 'feat(run): execute multi-phase pipelines (RunnableAgent.Phases)' (#19 ) from feat/kernel-phases into main
steve
2026-06-29 19:52:51 +00:00
0dd2ced717
fix(run): address gadfly review of the phases PR
steve
2026-06-29 15:44:04 -04:00
30b79a330f
feat(run): execute multi-phase pipelines (RunnableAgent.Phases)
steve
2026-06-29 15:14:45 -04:00
b25a13ed4f
chore: repin gadfly reusable to @5007597 (structured findings + consensus + inline review)
2026-06-28 22:13:24 -04:00
add8f847a4
Merge pull request 'feat(run): InputFileStager seam — stage non-image attachments into the prompt' (#18 ) from feat/input-file-stager into main
steve
2026-06-28 18:19:28 +00:00
df4033f42e
fix(run): harden input-file staging per gadfly #18 validation pass
steve
2026-06-28 14:08:57 -04:00
1e65f4b6e5
fix(run): sanitize input-file names — path-traversal + prompt-injection hardening (gadfly #18 )
steve
2026-06-28 13:29:45 -04:00
2ef88f2a73
feat(run): InputFileStager seam — stage non-image attachments into the prompt
steve
2026-06-28 13:02:55 -04:00
7a5eebc468
fix(ci): restore valid adversarial-review.yml + pin gadfly reusable @7bc3c98 [skip ci]
2026-06-28 02:05:28 -04:00
7211ce227c
ci: pin gadfly reusable workflow to immutable sha (cache-bust @v1)
steve
2026-06-28 05:44:52 +00:00
f367796244
Merge pull request 'run: fold inv.Images into the initial user message (multimodal opening turn)' (#16 ) from feat/initial-images into main
steve
2026-06-28 05:16:46 +00:00
0acaa8c9a5
run: guard empty text part in runAgent + drop cross-repo doc ref (gadfly #16 )
steve
2026-06-28 01:11:15 -04:00
a35c176b42
run: fold inv.Images into the initial user message (multimodal opening turn)
steve
2026-06-28 00:37:53 -04:00
1cf46c9954
ci: track gadfly's v1 release tag instead of a pinned sha (#15 )
steve
2026-06-28 04:08:30 +00:00
56baac758d
ci: inherit gadfly's default swarm (slim caller, re-pin @b02b11d) (#14 )
steve
2026-06-28 02:48:25 +00:00
5779035722
Merge pull request 'ci: subscribe to gadfly's reusable review workflow (cloud + Claude Code, no local)' (#13 ) from ci/gadfly-reusable into main
steve
2026-06-28 01:43:42 +00:00
1a2a2364ec
security: scope forwarded secrets + pin gadfly reusable to an immutable sha
2026-06-27 21:18:59 -04:00
c08ce47fa6
ci: subscribe to gadfly's reusable review workflow (cloud + Claude Code, no local)
2026-06-27 20:03:00 -04:00
784d5d7ce4
run: PostRun detached ctx + panic-isolated Cleanup (gadfly #12 )
steve
2026-06-27 18:33:41 -04:00
4e179259de
run: wire SessionToolFactory + PostRun artifacts + AttachImages
steve
2026-06-27 18:13:16 -04:00
82a816ae29
ci(gadfly): trim pool to the strong 6 (drop m5/qwen3.6, gemma4, gpt-oss, kimi-k2.7)
steve
2026-06-27 18:06:36 -04:00
be4bbbcad5
run: fix statusFor — don't relabel a generic error / caller-cancel as timeout (gadfly #11 )
steve
2026-06-27 17:00:26 -04:00
390e6cf905
run: critic parity — fuller RecordStep + cause-carrying Kill (distinct status)
steve
2026-06-27 16:35:13 -04:00
1a1d5e417b
chore: go mod tidy (add missing go.sum entry; CI tidiness gate)
steve
2026-06-27 14:53:58 -04:00
f3bd43b726
ci(gadfly): drop the m1 reviewer (dead weight; keep m5)
steve
2026-06-27 14:41:14 -04:00
306d575c31
critic: overflow-guard maxSteps += RaiseStepsBy (gadfly 5-model convergence)
steve
2026-06-27 14:38:48 -04:00
4ba83ab905
run: critic can raise a run's step ceiling mid-flight (CriticHandle.MaxSteps)
steve
2026-06-27 14:16:03 -04:00
a103cc5e9f
ci(gadfly): 9-cloud panel @ 3 models x 3 lenses (9 concurrent)
steve
2026-06-27 12:17:24 -04:00
4d28cd6e2c
ci(gadfly): 4-cloud pool — add kimi-k2.7-code + deepseek-v4-pro, drop v4-flash
steve
2026-06-27 11:59:13 -04:00
dcaefff756
ci(gadfly): add M1/M5 Macs back to the reviewer pool (full fleet)
steve
2026-06-27 10:44:22 -04:00
97154395e6
C0b: document recordToolStart post-iteration timing (gadfly glm finding)
steve
2026-06-27 10:10:56 -04:00
4aa06f652e
C0b: address verified gadfly findings (panic-safety + test honesty)
steve
2026-06-27 10:09:22 -04:00
43b2471737
C0b: wire Critic + Delivery into run.Executor
steve
2026-06-27 10:00:05 -04:00
0c80679719
C0: address verified gadfly findings (trivial fixes)
steve
2026-06-27 09:53:11 -04:00
9d41987b0e
C0: wire Palette delegation into run.Executor (skill__/agent__ tools)
steve
2026-06-27 09:28:01 -04:00
e37cf415de
ci(gadfly): emit findings to gadfly-reports + bump image to sha-d7f364d
steve
2026-06-27 09:12:46 -04:00
a87e7d2c72
fix: address verified gadfly P5 findings (canary robustness)
steve
2026-06-27 00:34:01 -04:00
ea9475da54
P5: light-tier canary — gadfly-shaped reviewer on executus core
steve
2026-06-27 00:22:02 -04:00
dc2d4ec425
P4c: remaining batteries — checkpoint + schedule + critic
steve
2026-06-27 00:15:32 -04:00
c8559676ed
P4b: skill noun + contrib/store (SQLite for budget/persona/skill/audit)
steve
2026-06-27 00:15:00 -04:00
d82cef46b4
fix: address verified gadfly P4/#4 findings (audit/budget/persona)
steve
2026-06-26 23:44:34 -04:00
2260480c81
P4: persona noun — Agent + ToRunnable bridge + Memory store
steve
2026-06-26 22:24:18 -04:00
9116abcae2
P4: budget battery — DBBudget (rolling 7-day) over run.Budget
steve
2026-06-26 22:17:51 -04:00
4d2f85d139
P4: audit battery — run.Audit Sink + Writer + queryable Memory store
steve
2026-06-26 22:14:37 -04:00
d0bd3ec3d9
fix: address verified gadfly P3 review (3-cloud fleet)
steve
2026-06-26 22:31:59 -04:00
78e6858751
P3: store group — kv_* + file_* tools (agent memory)
steve
2026-06-26 22:06:46 -04:00
1e201550b3
P3: meta + primitive tool group (think/now/cite + classify/extract/summarize)
steve
2026-06-26 21:00:45 -04:00
df95425bb5
P3 (kickoff): generic tools/ library + end-to-end tool-using-agent test
steve
2026-06-26 20:54:28 -04:00
16ddd90914
ci(gadfly): new build sha-d0de034 + per-lens concurrency
steve
2026-06-26 22:57:14 -04:00
b35514dfaa
ci(gadfly): cloud-only fleet (3 models, drop local Macs)
steve
2026-06-26 22:01:07 -04:00
7b3da87c08
fix: address verified gadfly P2 findings (9 real of 18)
steve
2026-06-26 21:42:46 -04:00
dfbc5a42b9
P2: run.Executor — executus is runnable
steve
2026-06-26 20:45:10 -04:00
130c2bdfab
P2: move compactor -> compact/ + step instrumentation -> run/steps.go
steve
2026-06-26 20:36:32 -04:00
d9b44387f5
fix: address gadfly P1 review (3 low-risk findings)
steve
2026-06-26 20:32:58 -04:00
e0e2c0451a
ci: sync gadfly review config to mort's foreman-provider setup
steve
2026-06-26 20:22:26 -04:00
c732a677df
P2: define nil-safe run.Ports (the inversion spine)
steve
2026-06-26 20:17:26 -04:00
fa644f1826
P2 (foundation): run-loop mechanics + RunnableAgent DTO
steve
2026-06-26 19:58:20 -04:00
b424261aca
P1: model layer (convar->config inversion) + llmmeta
steve
2026-06-26 19:47:13 -04:00
741d7816ed
ci: add gadfly adversarial review on PRs (mirrors mort)
steve
2026-06-26 19:46:47 -04:00
dc28b63ad8
P1 (part 1): move skilltools core -> tool/ (clean, verbatim)
steve
2026-06-26 19:31:47 -04:00
d2c18ad5bb
ci: make tidy-clean check robust to a missing go.sum
steve
2026-06-26 19:20:47 -04:00
ca243a2d50
P0: stand up executus harness module above majordomo
steve
2026-06-26 19:18:37 -04:00
25feb63c00
Initial commit
steve
2026-06-26 23:04:54 +00:00
steve