 Jesse Vincent
|
ec014e7a7f
Bump evals submodule to merged superpowers-evals main (ac264b1)
|
3 months ago |
|
Jesse Vincent
|
eba16f6b91
Strict-cost spec: L2 recon n=2 (sonnet controller $6.68/$8.05, judgment clean, escalation points unstressed)
|
3 months ago |
|
Jesse Vincent
|
27788fdef9
Strict-cost spec: record batch A-E rung verdicts (L1 validated, L2 recon positive, L3 dead)
|
3 months ago |
|
Jesse Vincent
|
9a25a75bac
Spec: strict-cost SDD experiment ladder — judgment as co-invariant, plan-side crispness first
|
3 months ago |
|
Jesse Vincent
|
60fa4f6fc4
Record writing-plans micro-test result: resolved, no change needed
|
3 months ago |
|
Jesse Vincent
|
43a6ee23f7
Spec: record iterations 4-5 (variance honesty, structural fixes, final validated ranges)
|
3 months ago |
|
Jesse Vincent
|
fe90d6c469
Adopt audited positive phrasings: evidence rule leads positive; fix-report completeness as checklist
|
3 months ago |
|
Jesse Vincent
|
b81f35bb1e
Land eval-tuned combo: file handoffs, progress ledger, final-review package, REQUIRED model lines, reviewer risk budget
|
3 months ago |
|
Jesse Vincent
|
926096a1d7
Spec: positive-instruction redesign — audit results, micro-test method, writing-plans variants
|
3 months ago |
 Jesse Vincent
|
a995af2e24
Shared: unique review-package collateral names
|
3 months ago |
|
Jesse Vincent
|
d4dbf44162
Add review-package script; close fix-dispatch test gap
|
3 months ago |
|
Jesse Vincent
|
2434ef7f35
Describe the review design as current state, not as a delta
|
3 months ago |
|
Jesse Vincent
|
7cf78437e2
Spec: record iterations 2-3 results and final frozen-config matrix
|
3 months ago |
|
Jesse Vincent
|
e355795625
Hand reviewers the diff as a file, not a paste
|
3 months ago |
|
Jesse Vincent
|
29ee4e8e44
Reviewer skepticism covers the implementer's design rationales
|
3 months ago |
|
Jesse Vincent
|
28498a5cde
Make diff-pasting non-optional for task reviewer dispatch
|
3 months ago |
|
Jesse Vincent
|
5e2907fc4f
Close the Minor-severity escape hatch
|
3 months ago |
|
Jesse Vincent
|
e532f24df7
Spec: document cost iterations and the per-task review consolidation
|
3 months ago |
|
Jesse Vincent
|
e3c74fc1c9
Merge per-task reviews into one task reviewer (iteration 2)
|
3 months ago |
|
Jesse Vincent
|
3e3e1e701e
Cut review-cost drivers: turn-aware models, inline diffs, scoped evidence
|
3 months ago |
|
Jesse Vincent
|
853396e3ae
Add phrase-level pre-judging triggers to reviewer prompt rule
|
3 months ago |
|
Jesse Vincent
|
83d54f7ddd
Red Flags: never tell a reviewer what not to flag or pre-rate severity
|
3 months ago |
|
Jesse Vincent
|
c7900f1698
Close three review blind spots found by defect tracing
|
3 months ago |
|
Jesse Vincent
|
5cfdb75b94
Require explicit model on subagent dispatch
|
3 months ago |
|
Jesse Vincent
|
87825ff193
Forbid controllers pre-judging reviewer findings
|
3 months ago |
|
Jesse Vincent
|
09cb4d7361
Sync plan: escaped pre() pattern in Task 5 checks block
|
3 months ago |
|
Jesse Vincent
|
b3bb9a68d7
Fix plan doc: correct Task 1 grep expectation; sync Task 5 story block
|
3 months ago |
|
Jesse Vincent
|
71dc271a08
Sync plan's Task 5 blocks with review fixes
|
3 months ago |
|
Jesse Vincent
|
5aea3dca31
SDD controller: reviewer prompt budgets, ⚠️ handling, final-review pointer, model judgment
|
3 months ago |
|
Jesse Vincent
|
b3281c0227
Implementer prompt: re-run covering tests after fixing review findings
|
3 months ago |
