Jesse Vincent
|
b62616fc12
Release v6.0.2: stop shipping the evals submodule
|
hace 2 meses |
Jesse Vincent
|
a21956e48c
Release v6.0.1: Codex fixes
|
hace 2 meses |
Drew Ritter
|
29c0b1b7db
fix: read Codex plugin version from manifest (PRI-2240)
|
hace 2 meses |
Drew Ritter
|
cf32920d3a
fix: exclude repo metadata from Codex sync (PRI-1168)
|
hace 2 meses |
Jesse Vincent
|
284be5905e
Set v6.0.0 release date to 2026-06-16
|
hace 2 meses |
Jesse Vincent
|
77879bbb91
Bump evals submodule: unify per-agent bootstrap scenarios
|
hace 2 meses |
Jesse Vincent
|
c5a965101b
Bump version to 6.0.0
|
hace 2 meses |
Drew Ritter
|
b3ee712d3a
Add visual companion Prime Radiant branding
|
hace 2 meses |
Jesse Vincent
|
9c61797773
Draft Superpowers 6 release notes
|
hace 2 meses |
Jesse Vincent
|
b61b55013a
E37: pre-flight plan review — surface plan conflicts as one batched question before Task 1
|
hace 2 meses |
Jesse Vincent
|
be400204b3
Spec: L2b tested — opus structural win, sonnet transmission+attention gap (E35/E36); bump evals to 9919b27
|
hace 2 meses |
Jesse Vincent
|
530476fd00
L2b: plan-mandated defects are findings the human adjudicates
|
hace 2 meses |
Jesse Vincent
|
e97faafb5a
E27 stack: conditional impl tier + final-review tier pin + narration recipe + terse reviewer contract
|
hace 2 meses |
Jesse Vincent
|
cfe48c28ac
E03: cheapest-tier implementers when plan carries complete code (transcription hypothesis)
|
hace 2 meses |
Jesse Vincent
|
8bcefb12cb
Strict-cost spec: L2 final — died at gates; explicit escalation holds at sonnet, implicit adjudication does not
|
hace 2 meses |
Jesse Vincent
|
8e1262a3ba
writing-plans: task right-sizing, Global Constraints header, per-task Interfaces blocks
|
hace 3 meses |
Jesse Vincent
|
de4672b171
Constraints block is the reviewer's attention lens: copy spec verbatim, never improvise process rules
|
hace 2 meses |
Jesse Vincent
|
25192df30b
Strict-cost spec: L1 final — cost win re-attributed to complete-code plans; guidance owns fidelity/variance
|
hace 2 meses |
Jesse Vincent
|
f5e8df4252
Strict-cost spec: L2 recon n=2 (sonnet controller $6.68/$8.05, judgment clean, escalation points unstressed)
|
hace 3 meses |
Jesse Vincent
|
b5b3b5d99c
Strict-cost spec: record batch A-E rung verdicts (L1 validated, L2 recon positive, L3 dead)
|
hace 3 meses |
Jesse Vincent
|
30bbeefe89
Spec: strict-cost SDD experiment ladder — judgment as co-invariant, plan-side crispness first
|
hace 3 meses |
Jesse Vincent
|
d3dd1ecc7d
Record writing-plans micro-test result: resolved, no change needed
|
hace 3 meses |
Jesse Vincent
|
b2872a4a66
Spec: record iterations 4-5 (variance honesty, structural fixes, final validated ranges)
|
hace 3 meses |
Jesse Vincent
|
e9b88d05c8
Adopt audited positive phrasings: evidence rule leads positive; fix-report completeness as checklist
|
hace 3 meses |
Jesse Vincent
|
4298eac856
Land eval-tuned combo: file handoffs, progress ledger, final-review package, REQUIRED model lines, reviewer risk budget
|
hace 3 meses |
Jesse Vincent
|
69a00350ff
Spec: positive-instruction redesign — audit results, micro-test method, writing-plans variants
|
hace 3 meses |
Jesse Vincent
|
d7a8c07fe3
Shared: unique review-package collateral names
|
hace 3 meses |
Jesse Vincent
|
c30d822efe
Add review-package script; close fix-dispatch test gap
|
hace 3 meses |
Jesse Vincent
|
68c9ddb870
Describe the review design as current state, not as a delta
|
hace 3 meses |
Jesse Vincent
|
aa80399355
Spec: record iterations 2-3 results and final frozen-config matrix
|
hace 3 meses |